The AI Trade
165 bullish · 133 bearish · 81 neutral · exact words, dated and sourced
+161Is the AI buildout durable demand or an overbuild? The capex, the losses, and who blinks first.
The count, last 30 days: 44 bullish, 24 bearish. The tape leans bullish. A count, not a verdict.
The tape vs the talk
The line is Nasdaq, last 90 days. Each dot is a claim as it landed on the tape: filled is bullish, outlined is bearish. Tap a dot to read it. Where the tape went after is context, not a verdict.
The tape
Jul 13
Jul 20
Jul 27
Aug 03
Aug 10
Sep 07
bullishbearishclaims filed per week
What the people who move markets are saying about this, in their own words. These are views on the record, not scored bets: when a call here grows a number and a deadline, it moves to the scoreboard. Newest first.
Neutral
Matt Bloxham tells the anchor a DOJ antitrust finding that unwinds or penalises the Nvidia-Groq $20bn deal is a low-probability though possible risk.
So I think for me personally, that would be a kind of a low publicity risk
Bullish
Matt Bloxham tells the anchor TSMC's demand is not slowing, it cannot build capacity fast enough, and the chip demand-supply imbalance will last multiple years.
They recently upgraded their full year guidance and they're continuing to struggle to kind of build enough capacity to meet that demand. I think it's going to be a multi year imbalance between demand and supply.
Bearish
Matt Bloxham tells the anchor Oracle faces open questions on delivering its data-centre build-out and funding it, having raised heavy debt with a potentially dilutive equity raise pending.
And if they can fund it, they've raised a lot of debt. There's an equity raise out there too, which is potentially dilutive. So I think that there's a question there about the funding as well.
Neutral
Jack Caffrey tells the anchor that because AI capex is increasingly debt-funded, the US equity market is shifting from an earnings-driven market to a credit-availability market.
And so to some extent, we become away from an earnings market over time towards becoming more of a credit availability market.
Bullish
Jack Caffrey tells the anchor he owns energy as the historical hedge against high inflation, which is what would hurt the AI trade.
It would be high inflation, and so I think you wind up protecting against our inflation by owning energy, which has historically been the source
Bullish
Tom Lee says hyperscaler debt issuance is not a sign the AI boom ends badly, since every growth industry was debt-financed.
if someone says that that's a sign that the A.I. bubble is going to end badly I think then they're not really good students history
Bullish
Tom Lee wants to own AI bottlenecks (Nvidia, semis, memory, energy, power) and things that grow with AI units.
So I think it makes sense to say hey I want to own bottlenecks. I want to own things that grow with A.I. which is units
Bullish
Spencer Cox (Utah governor) says data centers will get built because the US is in an AI arms race with China, while criticizing how hyperscalers handled Utah
What I've said is they're going to be built. We're in an arms race with China. We need American models to win, and we can do this the right way.
Nasdaq since this was said: ▼ -0.7% · 26,253 → 26,082 · as of 2026-09-10
Bullish
Cox says chips will get more efficient and use less energy and water, so data-center environmental objections can be solved
The chips are going to be more efficient. They're going to use less energy, less water. So we can solve those problems.
Nasdaq since this was said: ▼ -0.7% · 26,253 → 26,082 · as of 2026-09-10
Bullish
Marc Ganzi (DigitalBridge CEO) says AI infrastructure capex of ~$900B/yr shows no sign of abating
We're spending almost $900 billion a year in CapEx and we don't see any of that abating anytime soon.
Nasdaq since this was said: ▼ -0.7% · 26,253 → 26,082 · as of 2026-09-10
Bullish
Ganzi argues Nvidia's circular investments differ from the dot-com fiber bust because leases are credit-wrapped hard equipment, not synthetic
This is a little bit structurally different because what's happening is the chip makers themselves aren't standing on the hook for the proprietary lease.
Nasdaq since this was said: ▼ -0.7% · 26,253 → 26,082 · as of 2026-09-10
Neutral
Ganzi states DigitalBridge has returned about $11B to investors from its early data-center bets (company milestone)
We've returned about $11 billion to investors because we were the first ones there.
Nasdaq since this was said: ▼ -0.7% · 26,253 → 26,082 · as of 2026-09-10
Neutral
Gave says hyperscaler debt needs to fund AI capex this year equal ~70% of the US budget deficit
their debt needs to fund the capital spending boom is roughly equivalent to 70 percent of the U.S. budget deficit.
Nasdaq since this was said: ▼ -0.7% · 26,253 → 26,082 · as of 2026-09-10
Bearish
Gave says the AI capex boom itself is pushing the price of capital and bond yields up by adding another deficit's worth of capital demand
So all of a sudden, it's like we've added another U.S. budget deficit into the balance, where, you know, that's more demand for capital. So everybody's competing for capital, the price of capital goes up.
Nasdaq since this was said: ▼ -0.7% · 26,253 → 26,082 · as of 2026-09-10
Bullish
Heath says copper is the unpriced binding constraint of the AI buildout because AI is a power problem before a chip problem
Copper is really, I think the part that nobody's pricing within the AI story. Like AI is really a power problem before it's a chip problem. And the power runs on copper, right? Everyone is watching AI trade through chips and the binding constraint is really copper.
Copper since this was said: ▼ -3.9% · 6.80 → 6.54 · as of 2026-09-11
Bullish
Heath says copper supply cannot respond to price because a major mine takes ~18 years versus 18 months for a data center
But the reality is you can build a data setter in 18 months. To build a major copper mine, it takes 18 years. So price is not going to change that.
Nasdaq since this was said: ▼ -0.7% · 26,253 → 26,082 · as of 2026-09-10
Neutral
Bradley Tusk tells CNBC that public anxiety plus real-world AI risk means government must now step up and regulate AI.
this is why we have regulation, and this is when government needs to step up and do their jobs.
Nasdaq since this was said: ▼ -0.7% · 26,253 → 26,082 · as of 2026-09-10
Bullish
Weiss says the AI trade is what keeps the market near highs and it is not going away, even as he is short the indexes
it's the ai trade so if that goes away then the bottom falls out but that's not going away
Nasdaq since this was said: ▼ -0.7% · 26,253 → 26,082 · as of 2026-09-10
Bullish
Wheeler thinks the AI build-out ultimately comes to Europe, whose ~10% share of the build-out could rise to 20-25%, bringing more local issuance.
we do think ultimately that comes to Europe and Europe today only has call it 10 percent of the A.I. build out. But we could see that get to 20 25 percent.
Nasdaq since this was said: ▼ -1.3% · 26,421 → 26,082 · as of 2026-09-10
Neutral
Campling suspects Anthropic walked from the ~$6bn Descartes deal because it can do the chip-efficiency work in-house.
I got a suspicion that maybe entropic can do this themselves now and don't need to spend six billion dollars on buying another company.
Nasdaq since this was said: ▼ -1.3% · 26,421 → 26,082 · as of 2026-09-10
Bearish
Peter Schiff says AI capex and hoped-for AI windfall profits are the only thing keeping a bid under cap-weighted US indexes; ex-AI the market is not doing well.
I think the US stock market has been held up by AI and all of the capex and the expectations of, you know, AI windfall profits, I think are keeping a bid in the US stock market.
S&P 500 since this was said: ▼ -1.1% · 7,674 → 7,592 · as of 2026-09-10
Bearish
Peter Schiff says the Mag 7 are far too popular and carry worse P/Es than the Asian tech names he owns; he is underweight tech.
And those MAG 7 stocks are just way too popular, I think.
Nasdaq since this was said: ▼ -1.3% · 26,421 → 26,082 · as of 2026-09-10
Bearish
Peter Schiff says American tech is in a bubble.
I think tech is a bubble in general. It's in a bubble.
Nasdaq since this was said: ▼ -1.3% · 26,421 → 26,082 · as of 2026-09-10
Bearish
Jeff Booth says high-cost data centers and Bitcoin miners paying 5-7c/kWh will go broke as mining migrates to 2.5c run-of-river and cheaper solar power.
the river run power is at two and a half cents and data centers here are paying five and six, seven cents. They're going to go broke.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Bearish
Jeff Booth says that, like Bitcoin treasury companies, everybody in AI today is buying the top and such late buying collapses on them.
Which would be normal, and for every AI today, everybody's buying the top, right? Because in the past they made all this money and they buy the top and it collapses on them.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Bullish
Sunny Banja (Bloomberg equity strategist) tells The Asia Trade that DeepSeek's price hike is the start of a trend: Chinese AI model prices will rise as the labs move to monetization over the coming months and years.
And we see hikes are probably a trend that we're going to see across all of these Chinese models over the coming months and years
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Bullish
Anthony Stevens (Bloomberg markets reporter) says demand for tech debt is enormous and expects AMD's $4.75bn bond deal to price tight, following Intel's well-received block.
So there is an incredible amount of demand for tech debt.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Bullish
Anthony Stevens (Bloomberg) argues China's AI buildout is a healthier ecosystem than past overcapacity waves because firms have margins, full order books and reinvest rationally, which can lift the broader Chinese economy.
So this seems to me a much more healthy ecosystem where people are actually making money, as Sunny says, and reinvesting that money in a rational way to make more money, frankly.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Bullish
Minmin Low (Bloomberg China correspondent) argues China has an AI adoption advantage over the US because Chinese users are more willing to experiment, even if China lags in chips.
And here is where China perhaps has an advantage over the U.S. because Chinese users are far more willing to experiment with new products.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Bullish
Julian Emanuel of Evercore tells Bloomberg he has not cut US equity exposure and remains overweight technology, preferring to hedge rather than sell.
And look, we have been overweight the technology theme the entirety of this bull market.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Bearish
Adam Farstrup (Schroders) tells CNBC to stay long US tech but complement it, because single-factor risk in the AI trade is very significant from here.
But think about how you can complement that because that risk within the trade of that one AI trade is very significant from here.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Neutral
Jurrien Timmer tells Bloomberg that about $70bn of US and Korean money has chased single-name levered semiconductor ETFs, creating pockets of crowded trade.
By my count between the U.S. and Korea about 70 billion dollars has been chasing that space. So there are pockets of crowded trade.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Bullish
Jurrien Timmer tells Bloomberg that demand for AI chips remains well ahead of supply despite crowded positioning.
overall you know from what we are seeing the demand for chips is still well ahead of the supply.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Neutral
Jurrien Timmer tells Bloomberg that roughly $500bn of investment-grade bond issuance this year and another $500bn next year will finance the AI buildout, raising a crowding-out question for Treasuries.
there's about half a trillion coming in investment grade bond issuance this year. Another half a trillion next year.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Neutral
Kim Forrest of Bokeh Capital tells Bloomberg the AI flywheel linking back to Nvidia's $500bn financing pool is concerning, though other winners have emerged this year.
Yeah it is concerning. Although I have to say I felt really good earlier this year that we've had some other winners other than Nvidia.
Nvidia since this was said: ▼ -3.0% · 225.2 → 218.4 · as of 2026-09-10
Neutral
Kim Forrest of Bokeh Capital tells Bloomberg LLMs will become a delivery layer over smaller targeted models rather than taking over the world.
I don't think LLMs take over the world. I think LLMs are going to be maybe the delivery mechanism but there's going to be other models in the back that have nothing to do with large language models.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Neutral
Kim Forrest of Bokeh Capital tells Bloomberg the AI compute buildout may be too much or may be repurposed for smaller targeted models; she is unsure.
that is going to use different kinds of compute. So are we building too much. Maybe. But maybe it's going to be turned into what I think are those smaller targeted models that are going to answer specific questions.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Neutral
James Kisner says investors are rightly no longer accepting hyperscaler AI capex on faith and are demanding evidence of payback.
So I think folks are appropriately cautious that we're not just spending with worthless abandon here, that we're going to see return on those dollars. So again, early days, but I think investors are becoming more discerning, broadly speaking.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Neutral
James Kisner says circularity in Nvidia financing its customers (OpenAI, neoclouds) is an appropriate concern that only resolves if revenue materializes.
I mean, I think it's, you know, it's an appropriate concern. You know, there's been a lot of talk about circularity in the marketplace.
Nvidia since this was said: ▼ -3.0% · 225.2 → 218.4 · as of 2026-09-10
Bullish
Gavin Baker tells Calacanis AI usage broadly accelerated in July 2026 even as AI-related public stocks traded poorly.
I think AI broadly accelerated in the month of July, which was kind of crazy to contrast with the public stock market action.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Bullish
David Sacks tells Calacanis demand for AI tokens will keep growing exponentially and the TAM is big enough; the constraint is physical supply.
I do think the TAM is big enough. I think the demand is going to be there.
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Bullish
Gavin Baker tells Calacanis Nvidia's bank/PE compute-financing platform is smart and good for everyone because top asset managers are validating GPUs as a financeable asset.
And I think that's smart. That's good for everyone.
Nvidia since this was said: ▼ -3.0% · 225.2 → 218.4 · as of 2026-09-10
Bullish
Gavin Baker tells Calacanis Nvidia GPUs have long useful lives: CoreWeave is renting 2020 Ampere chips profitably into 2029, implying about nine years.
an ampere that came out in 2020 is going to have a nine-year life
Nvidia since this was said: ▼ -3.0% · 225.2 → 218.4 · as of 2026-09-10
Bullish
Jason Calacanis tells Baker and Sacks US firms will spend 5-10% of employee salaries on AI tokens, implying roughly $1T (or at least $500B) of US AI spend, so there is no bubble.
If you put those numbers together, you're looking at a trillion dollars in AI spend just in the United States
Nasdaq since this was said: ▼ -2.4% · 26,729 → 26,082 · as of 2026-09-10
Bullish
Chey Tae-won (SK Group chairman) expects total memory volume to grow to almost five times current production capacity within 10 years, justifying the $720bn expansion.
So what I really see is that the memory increase in the total volume is within 10 years, I guess, almost five times current, the production capacity. So demand is enormous.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bullish
Chey Tae-won says 2027 will be the hardest year for memory supply because capacity cannot expand in time, while demand roughly doubles from this year.
So probably next year is at the hardest year because all of the memory companies are not actually prepared to expand their capacity yet. But still demand is growing. I think that the demand is growing twice than this year.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Neutral
Chey Tae-won says memory prices have risen 40-50%, a 'chip inflation' problem for PC, smartphone and auto makers.
So yeah, memory prices have gone up 40 50 percent.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Neutral
Chey Tae-won reiterates SK Hynix will double capacity within five years and says customers still want more.
So I already announced that within five years, we're going to double up our capacity. But most of the customer, that's not enough. So we need more.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Neutral
Chey Tae-won concedes some oversupply risk in memory but says SK Hynix expands only against confirmed demand and long-term agreements.
There is some risk on the oversupply. But we already noticed that and we carefully confirm that the world demand and the inner futures.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bullish
Chey Tae-won thinks AI agents make this memory boom a structural change with a much longer cycle rather than the usual boom-bust, while admitting nobody can see the future.
But I feel that this is not a normal things and what we used to be the in the memory business. So this is a kind of turning point for us as what I see that do I see that some structural change.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bullish
Siegel tells CNBC non-tech S&P 500 margins can expand dramatically if firms use AI to cut costs.
margins overall on the S&P are at all-time highs. But if you take out the tech communication, they've been relatively stable. But they can improve dramatically if they can use what AI is designed for, and that is to reduce costs.
S&P 500 since this was said: ▼ -2.7% · 7,799 → 7,592 · as of 2026-09-10
Neutral
Dan Ives says the biggest risk to the AI revolution is politics blocking data-center buildouts (e.g. New York's moratorium), not capex or use cases.
So I actually think the biggest sort of risk, when you think about what could happen to AI revolution, it's not cap backs. It's not use cases. The biggest risk is just the political piece gets in the way in terms of the data center build outs.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Neutral
Dan Ives says AI chip demand currently exceeds supply by roughly 13-14 to 1.
So I view it as, right now, demand the supply for chips, 13, 14 to 1, demand the supply.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bullish
Dan Ives says investors are still way underestimating the scale and scope of the AI revolution and the AI trade.
and it's always been our view-- that investors are way underestimating the scale and scope of the AI revolution.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bullish
Dan Ives says AI chip supply will not reach equilibrium with demand until probably 2028 or early 2029.
Well, equilibrium, you won't hit till probably 2028, maybe early 2029.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bearish
Ed Zitron (EZPR) says that by his reporting 69% of Microsoft Intelligent Cloud's 2025 growth came from OpenAI, ex-OpenAI growth was only 8%, and calls the AI sales narrative 'kind of a lie'.
So in calendar year 2025, according to my own reporting about open AI's numbers, 69% of the year over year growth of Microsoft Intelligent Cloud segment was actually from open AI. Without that, it would have only grown 8% year over year, which is barely beating inflation.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bullish
Dan Ives says betting against the AI trade is betting against a capex cycle that continues to accelerate.
And that's why I think those that bet against this, it's betting against a CapEx cycle that actually continues to accelerate.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Neutral
Rick Davis (Stone Court Capital, Bloomberg politics contributor) says Virginia public support for data centers has flipped from over 60% to about 30% over water and power-bill fears.
over 60% of the people of Virginia supported that now it's in reverse now it's only 30% of the people in Virginia support it
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Neutral
Jeanne Zaino (Bloomberg politics contributor) says the five largest hyperscalers are on track to spend over $1 trillion on AI infrastructure across last year and this year.
if you look at the five biggest hyperscalers that includes Amazon that includes Google and Microsoft that they're on track to spend over one trillion dollars on AI infrastructure from last year to this year
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bearish
Rick Davis predicts tighter local, state and federal regulation of data centers ('bring your own power') is coming, while urging the issue not be made partisan.
I don't think there's any question that the country's in a phase where you're going to tighten regulations around these
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bearish
Jeanne Zaino argues the US needs a pause on data-center construction, praising Hochul's moratorium and warning earth-bound infrastructure could be stranded.
we do need a pause we do need some kind of time out where we think about what the impact of all of this is on our lives and our communities
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bearish
Roger McNamee says holders of AI-exposed stocks such as semiconductors should question their assumptions because some of them are no longer sound.
Because there are real issues that if I own the stocks today, I'd be sitting there and asking a question. Are all my assumptions still sound? Because I think a few of them are not.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bearish
Roger McNamee says the assumptions that current LLM technology will be economically viable and transform the economy for the better are deeply flawed.
Again, it assumes that the current technology is going to produce something economically viable that will transform the economy for the better. I think both of those assumptions are deeply flawed.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bearish
Megan Horneman (Verdence Capital Advisors) tells the anchor that AI capex spending will not continue at this pace forever, though she cannot say when it will turn or peak.
We don't know when this is gonna turn, we don't know exactly when this is gonna peak, but the spending pace isn't gonna continue forever.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Neutral
Chad Dillard (Bernstein) tells CNBC the data-center project pipeline has tripled to ~300 GW, nearly as much power as the US consumes.
the pipeline of projects have increased by 3x for data centers just to put some numbers around it.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Neutral
Chad Dillard (Bernstein) tells CNBC stranded/cancelled data-center capacity has held steady at ~11% despite the Texas and New York hookup moratoriums.
11% of stranded capacity that's, you know, been existing over the last, you know, couple of months, that really hasn't changed.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bullish
Mike Pyle (BlackRock) tells CNBC he sees no slowdown in the AI capex wave and considers the environment still underbuilt, not overbuilt, with no deterioration in capex expectations this earnings season.
we don't see any slowdown of this incredible CAPEX wave that looks to us as if this is an environment that's still underbuilt, not overbuilt
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bullish
Clem Chambers says copper will double within the next year on electrification/AI demand and supply shortage, while gold will not double.
I mean, I think copper, copper, I look at that chart and I go, it's going to double in the next year. And, you know, gold ain't going to double.
Copper since this was said: ▼ -0.9% · 6.59 → 6.54 · as of 2026-09-11
Neutral
Clem Chambers would not buy semiconductors after the August rebound; he sees a possible bull-trap shape and would only buy on a breakout above the range.
No, no, I would be watching that top. And if it breaks through the through that that range and starts heading up, I would buy it, but it could easily break down and drop a lot.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bearish
Rohan Goswami (Semafor) tells Prof G Markets that CME's AI compute futures will never be a real market because Nvidia alone controls supply and price, like Standard Oil.
it is not ever, in my opinion, going to be a real market because it would be analogous to Standard Oil
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bearish
Rohan Goswami says Nvidia's $500bn financing MOU spreads risk among more Wall Street players but does not reduce the key-man risk concentrated in Nvidia, citing Lucent.
well, there's no really avoiding this risk. Just because there are more players involved doesn't mean the actual risk has gone down.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Bearish
Ed Elson (Prof G Markets) argues that a dozen OpenAI executive departures, a delayed IPO and mounting losses say 'nothing good' about a stock market that has become a bet on OpenAI, though he is not recommending selling.
bailing on OpenAI say about the stock market? And the answer is probably nothing good.
Nasdaq since this was said: ▼ -2.7% · 26,803 → 26,082 · as of 2026-09-10
Neutral
Neil Shah (Counterpoint Research) tells Bloomberg that top hyperscalers are spending $700-800bn a year on AI capex, so Nvidia's $500bn financing plan spread over years is only 20-30% of that and mostly relevant to tier-2/3 players.
it's almost reaching 700 to 800 billion dollars per year for the top hyperscalers. And when you look at this 500 billion dollar fund
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bullish
Neil Shah says there is a lot of real AI compute demand and adoption is only scratching the surface, with maybe 5-10% of people and businesses using AI so far.
Absolutely there is a lot of demand we can see and we are just scratching the surface right now.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Neutral
Dan Niles (Niles Investment Management) says AI data centers are legitimately becoming a securitized asset class that will make financing easier, though Nvidia's flat stock shows investors are not yet convinced.
is AI data centers an asset class? Absolutely. Is there a lot of money being spent? 100%.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bullish
Dan Niles thinks the AI bubble has at least another year to build, with very good returns, before it eventually bursts, by analogy with 1999-2000.
So I think we've got at least another year in this bubble building out. And I think the returns are going to be very, very good.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bearish
Dan Niles expects about another year of AI buildout and then, at some point, a bust in which the new financing schemes will have problems.
So I think we got about another year in this build out. And then at some point you're going to get a bust
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Neutral
Dan Niles says the six biggest hyperscalers have over $1 trillion of off-balance-sheet financing.
There's over one trillion dollars in off balance sheet financing just looking at the six biggest guys out there.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Neutral
Dan Niles says hyperscaler capex growth is expected to accelerate to almost 100% year over year in the September quarter, supporting Nvidia's valuation.
expected to accelerate even further in the September quarter, almost 100 percent year over year growth
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bullish
Neil Campling (Bloomberg) tells Francine Lacqua that AI compute demand is very strong after CoreWeave's results.
I think the demand clearly is very, very strong.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Neutral
Campling tells Lacqua CoreWeave's revenue is ~80% concentrated in three hyperscalers and its costs are Nvidia-heavy (neutral description).
23 percent of their cost is in video chips, 80 percent of their revenues is coming from the three massive hyperscalers.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Neutral
Campling tells Lacqua CoreWeave's pre-market jump is partly short covering given ~18% short interest.
there is actually a very large base of short sellers in the stock, about 18 percent of outstanding on the short.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bullish
Campling tells Lacqua the Wall Street/PE financing arrangement reduces the circular-funding risk that had widened Nvidia CDS (mildly positive for Nvidia).
And so you're diluting part of that risk, if you like, that is less around a circular funding.
Nvidia since this was said: ▼ -2.6% · 224.1 → 218.4 · as of 2026-09-10
Bullish
Matt Bloxham (Bloomberg Intelligence) tells Lacqua memory-chip price pressure is still there and will show up in company numbers in H2 2026.
But I think that that that whole kind of upward pressure and memory pricing is still very much there. And it's still going to kind of populate through the market and in numbers in the second half of the year.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Neutral
Dibo tells Lacqua hyperscaler capex will exceed operating cash flow this year and next, which is why the market worries about AI monetization.
And importantly the capex will exceed the cash flow from operations for these companies both this year and next.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bullish
Mitchell Green (Lead Edge Capital) tells CNBC that AI will change the world and probably be bigger than the internet.
the answer is AI is going to change the world. It is going to revolutionize things. It will probably be bigger than the internet.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bearish
Jay Goldberg of Seaport tells Prof G Markets that Nvidia's $500bn financing plan looks like circular financing, which makes investors uncomfortable though not enough to sell.
this certainly looks like it's a big piece of circular financing going on here because Nvidia is providing money to customers to buy Nvidia parts.
Nvidia since this was said: ▼ -2.6% · 224.1 → 218.4 · as of 2026-09-10
Bearish
Jay Goldberg of Seaport tells Prof G Markets that pro-cyclical debt will amplify Nvidia's pain when the AI cycle turns and neoclouds put unused capacity back.
The problem is when the cycle turns and the cycle always turns, this kind of thing will amplify the pain on the downside, right?
Nvidia since this was said: ▼ -2.6% · 224.1 → 218.4 · as of 2026-09-10
Neutral
Jay Goldberg of Seaport tells Prof G Markets that Nvidia had $30bn of off-balance-sheet compute service agreements as of last quarter.
NVIDIA had $30 billion of what they call compute service agreements in place. It's not on the books, it's not on the balance sheet
Nvidia since this was said: ▼ -2.6% · 224.1 → 218.4 · as of 2026-09-10
Bullish
Tiffany McGhee (Pivotal Advisors) tells Bloomberg she remains very constructive on AI now and for the future, even as the story shifts from spending excitement to scrutiny of returns.
we're certainly just to be clear, certainly very constructive on on AI and not just now, but for the future.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bullish
Bob Greifeld (Cornerstone Financial Tech Management, ex-Nasdaq CEO) tells CNBC Nvidia's stock should go up because widening the financing circle to third parties lessens Nvidia's own vendor-financing risk and lets the market set rates.
when you think about it the stock should go up because in a certain way they've lessened their rights right so they started with situation where they're providing product to the customer costs paying for it then they're in the business of vendor financing.
Nvidia since this was said: ▼ -2.6% · 224.1 → 218.4 · as of 2026-09-10
Bearish
Bob Greifeld (Cornerstone Financial Tech Management) tells CNBC that with more parties in Nvidia's financing circle there is leverage at every point, so nobody really knows how much leverage is in the AI financing system.
So one there's more people in the circle and this points of leverage in every part of the circle so we don't really know how much leveraging is in the system as we go through that.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bearish
Bob Greifeld (Cornerstone Financial Tech Management) tells CNBC that because LLMs were built for effectiveness not efficiency, the compute needed per LLM task will fall roughly a hundredfold over the next decade, a risk for lending against data-center equipment.
so i would think that the compute power required to do a given task in an llm is going to go down a hundredfold
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bullish
Bob Greifeld (Cornerstone Financial Tech Management) tells CNBC he agrees that older GPUs such as 2020-era A100s have longer useful lives than people assumed a couple of months ago.
It does i'm in violent agreement with that i believe the useful life of these computers be longer than people were talking about you know just a couple months ago
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bearish
Thomas Hayes (Great Hill Capital) tells David Lin the August rebound in semiconductors/memory is a dead-cat bounce he would sell, expecting it to fade around the third/fourth week of August as institutions use it to raise cash.
I would be a seller of this move, it's a dead cat bounce and that's why you see these waterfalls like you saw in 2000, like you saw in 2022
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Neutral
Thomas Hayes (Great Hill Capital) tells David Lin that AI-related exposure in leveraged ETFs has jumped from 26% to 58% in the last couple of years, a warning sign of leverage in the AI trade.
AI related exposure to leverage ETFs now has jumped from 26% to 58% in the last couple of years
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Neutral
Thomas Hayes (Great Hill Capital) tells David Lin that Whale Rock (-21.7%), Lone Pine (-27%) and Altimeter (-11%) lost heavily in July, evidence that institutions must deleverage into the bounce.
whale rock down 21.7% in the month of July, lone pine, 27%. Altimeter, okay, down 11% in the month of July.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Neutral
Thomas Hayes (Great Hill Capital) tells David Lin that tech revenue growth (15%) and earnings expectations have not been this high since Q4 2021, which in his view preceded the 2022 tech correction.
have not seen revenue growth, or earnings expectations at levels this high, you know, 15% revenue growth since q4 of 2021.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bullish
Thomas Hayes (Great Hill Capital) tells David Lin he is not bearish on the long-term AI thesis (mid innings) but the return on investment will show up much later than the market assumes.
we're not generally bearish on the AI thesis. We think we're mid innings, but the return on investment is going to show up much later.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bearish
Thomas Hayes (Great Hill Capital) tells David Lin that hedge funds hit in July face redemption requests and will sell AI/semiconductor stocks into the dead-cat bounce in coming weeks.
These are guys that are going to wait for the dead cat bounces to create the liquidity that they need to meet the redemption requests. And that's coming to a theater near you in coming weeks.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Neutral
Thomas Hayes (Great Hill Capital) tells David Lin the hyperscalers are free-cash-flow negative and cannot buy back stock, with FCF not expected to inflect until late 2027 or early 2028.
That's not expected to inflect until or start to turn up until the late 27, early 28. That's predicated on we hope and expect that we'll start to see a return on invested capital.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bullish
Thomas Hayes (Great Hill Capital) tells David Lin that AI investment will eventually show up in productivity, but only in another 18-24 months, and the despondency phase before then will be the time to buy the AI trade.
I'm of the camp that the answer will be yes, but it will take another 18 to 24 months. And there will be a point in time during that period where people stop believing that the return will come before the bill comes due. And that will be the time to buy the AI trade.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bullish
Thomas Hayes (Great Hill Capital) tells David Lin that Alphabet's $85bn raise looks like a bad use of capital on short-term results, but he is in the camp that AI capex pays off eventually, just on a longer runway than the market expects.
And I'm kind of in that camp. I just think that the runway to that brighter day is a bit longer than the market thinks it's going to be right now.
Nasdaq since this was said: ▼ -1.9% · 26,588 → 26,082 · as of 2026-09-10
Bullish
Anita Krishna Gupta of Wealthbrix tells Bloomberg the AI-led rally is not going away, with leadership rotating to cloud names like Alphabet, Microsoft and Amazon.
that's not going away. It's just that different parts of that A.I. bucket are going to perform or not perform. So what's performed more recently is anything related to cloud services.
S&P 500 since this was said: ▼ -1.8% · 7,728 → 7,592 · as of 2026-09-10
Bullish
David Sacks tells Fox Business the US is roughly six to 12 months ahead of China in frontier AI models.
We are way ahead of China. I think we're probably about six to 12 months ahead of China in terms of the models.
Nasdaq since this was said: ▼ -1.4% · 26,445 → 26,082 · as of 2026-09-10
Bullish
David Sacks tells Fox Business the US will keep its AI lead as long as companies are not burdened with unnecessary bureaucratic hoops.
So as long as we don't subject our companies to a lot of unnecessary bureaucratic hoops, we will stay in the lead in this AI race.
Nasdaq since this was said: ▼ -1.4% · 26,445 → 26,082 · as of 2026-09-10
Neutral
David Sacks tells Fox Business AI capex is $800bn this year and expected to exceed $1.4tn, or 3% of GDP, next year.
So CAPEX, which is basically just investment in the hardware layer and construction of data centers, is expected to be 3% of GDP next year. It's $800 billion this year. It's over $1.4 trillion next year.
Nasdaq since this was said: ▼ -1.4% · 26,445 → 26,082 · as of 2026-09-10
Bullish
David Sacks tells Fox Business the software-side productivity boom from AI is only just getting started.
The boom on the software side, which is the productivity gains from all these companies and these consumers becoming more productive because they're using the software, that is just getting started.
Nasdaq since this was said: ▼ -1.4% · 26,445 → 26,082 · as of 2026-09-10
Bullish
Julian Emanuel tells CNBC the recent AI financing deals show the AI buildout is alive and well, with the caveat that it depends on capital-markets access.
you see that the AI revolution is very much alive and well, but it's dependent on capital markets.
Nasdaq since this was said: ▼ -1.4% · 26,445 → 26,082 · as of 2026-09-10
Neutral
Neil Campling tells Bonnie Quinn it is unusual and a question mark ahead of earnings that Nvidia's CDS now trades wider than the hyperscalers' despite its strong balance sheet.
this is quite an unusual move to see NVIDIA's CDS go above some of the other hyperscalers, I think.
Nvidia since this was said: ▲ +0.4% · 217.5 → 218.4 · as of 2026-09-10
Neutral
Laura Cooper tells Bonnie Quinn hyperscaler debt has doubled to about 4% of the US investment-grade market.
we've already seen hyperscaler debt as a share of the U.S. investor grade market double to about 4 percent.
Nasdaq since this was said: ▼ -1.4% · 26,445 → 26,082 · as of 2026-09-10
Neutral
Laura Cooper tells Bonnie Quinn that on ~$2tn capex estimates, hyperscalers could become a quarter or more of the US investment-grade market.
these hyperscalers could account for anywhere upwards of one quarter of the U.S. debt market in terms of investment grades.
Nasdaq since this was said: ▼ -1.4% · 26,445 → 26,082 · as of 2026-09-10
Bearish
Laura Cooper tells Bonnie Quinn Nuveen prefers high-quality hyperscalers and AI-adjacent old-economy credit over pure-play AI software, which is more vulnerable to disruption.
some of that pure play A.I. software that is more vulnerable
Nasdaq since this was said: ▼ -1.4% · 26,445 → 26,082 · as of 2026-09-10
Bullish
Matt Bloxham (Bloomberg Intelligence) tells Bonnie Quinn the AI economy is hungry for capital and, judging by the Nvidia $500bn platform, Intel and Google raises, the capital is there.
So I think generally that the AI economy is hungry for capital and the capital seems to be there.
Nasdaq since this was said: ▼ -1.4% · 26,445 → 26,082 · as of 2026-09-10
Bearish
Matt Bloxham tells Bonnie Quinn the Nvidia financing platform deepens AI's circular financing and raises questions about systemic risk down the line.
And it just kind of adds to this complex circular economy that we have building around AI and potentially I think raises some questions about the possibility of systemic risk somewhere down the line.
Nasdaq since this was said: ▼ -1.4% · 26,445 → 26,082 · as of 2026-09-10
Bearish
Ian Rogers (Ledger) says an agentic future where people hand passwords, cards and identities to AI agents is a security nightmare, and recent hacks fit that pattern.
an agentic future where we give our passwords, our credit cards, and our identities to agents is a security nightmare.
Nasdaq since this was said: ▼ -1.4% · 26,445 → 26,082 · as of 2026-09-10
Bullish
Jeremy Siegel says tech margins are incredible and hyperscaler spending, which he puts at 40% of the S&P, can continue to do well even as workers struggle.
the hyper-scale spending is doing well it's forty percent of the s_n_p_ so that can continue to go do well
Nasdaq since this was said: ▼ -2.0% · 26,605 → 26,082 · as of 2026-09-10
Neutral
Mark Zandi of Moody's tells CNBC the AI capex boom is largely imported equipment and will not lift US growth above 2% or create many jobs, though productivity gains may follow.
not going to give us above 2% growth, and it's not going to create, you know, boatload of jobs.
Nasdaq since this was said: ▼ -2.0% · 26,605 → 26,082 · as of 2026-09-10
Bullish
Matt Bloxham tells Bonnie Quinn TSMC's 45% July sales jump shows near-term demand for AI compute and chips is incredibly strong.
near term demand for A.I. compute and all the chips that go into that is incredibly strong
Nasdaq since this was said: ▼ -2.0% · 26,605 → 26,082 · as of 2026-09-10
Neutral
Matt Bloxham tells Bonnie Quinn Apple's near-term access to CXMT chips is limited because CXMT prioritizes Chinese customers; it is a long-term diversification play.
their ability to tap into this company as a near term supply could be limited because that company is prioritizing local Chinese manufacturers
Nasdaq since this was said: ▼ -2.0% · 26,605 → 26,082 · as of 2026-09-10
Bearish
Horacio Rozanski tells CNBC the AI-driven cyber threat has entered a vertical acceleration phase and will intensify as the technology spreads.
And we now believe we're in this vertical part of the curve where the cyber AI threat is real, is here right now.
Nasdaq since this was said: ▼ -2.0% · 26,605 → 26,082 · as of 2026-09-10
Bearish
Horacio Rozanski tells CNBC the cost of sophisticated cyberattacks has fallen from millions to thousands and will fall further as Chinese models catch up to US frontier models.
And as the Chinese models begin to catch up to our frontier models in this area, it's going to get even cheaper and more widespread.
Nasdaq since this was said: ▼ -2.0% · 26,605 → 26,082 · as of 2026-09-10
Neutral
Horacio Rozanski tells CNBC adoption matters more than innovation in the US-China AI race, so over-regulation could hand China an edge.
And this is where China can get an edge if we over-regulate this field.
Nasdaq since this was said: ▼ -2.0% · 26,605 → 26,082 · as of 2026-09-10
Neutral
Bloomberg The China Show, Korea markets roundup; reporter's own analysis of SoftBank's role in the AI trade.
SoftBank is obviously the fundraiser of choice in the AI trade globally.
Nasdaq since this was said: ▼ -2.3% · 26,691 → 26,082 · as of 2026-09-10
Bearish
Prof G Markets w/ Ed Elson; none of the big AI spenders has articulated the business model.
But none of these companies is enunci ated what exactly the business model it is that they hope to deliver.
Nasdaq since this was said: ▼ -2.3% · 26,691 → 26,082 · as of 2026-09-10
Bearish
Prof G Markets; Big Tech now effectively manufacturers building AI capacity, judged like manufacturers.
They're building huge capacity for whatever, AI products and services.
Nasdaq since this was said: ▼ -2.3% · 26,691 → 26,082 · as of 2026-09-10
Bullish
Prof G Markets; cites Amazon's $225B 2026 capex as why Nvidia's quarter should be good; explicitly not joining Burry's bear call.
Amazon is spending $225 billion this year on CapEx? So until we
Nasdaq since this was said: ▼ -1.0% · 26,348 → 26,082 · as of 2026-09-10
Neutral
Prof G Markets; the mega-cap AI spenders flipped from cash machines to negative free cash flow; a major shift he flags without joining the bears.
you know, companies like Microsoft and Google and Amazon, you know, companies who once threw off cash like it was water, have negative cash flow.
Nasdaq since this was said: ▼ -1.0% · 26,348 → 26,082 · as of 2026-09-10
Bullish
Responding to Cramer's point that analysts on the call were too negative on data center AI numbers; Su says 'we are in an extraordinary time for AI compute right now.'
We just updated our overall market numbers. We see the market for overall computing and high performance and adaptive computing going up to over $2 trillion as we go through 2030.
S&P 500 since this was said: ▼ -1.7% · 7,724 → 7,592 · as of 2026-09-10
Neutral
Asked how enterprise IT budgets for next year will treat AI , cost discipline vs. pedal to the metal; Su says 'a little bit of both.'
the AI costs are getting so large that everybody has to manage just like you manage your employee costs, you have to manage your AI costs.
S&P 500 since this was said: ▼ -1.7% · 7,724 → 7,592 · as of 2026-09-10
Bullish
Hyperscalers shifting from 'panic'-feel capex talk to proof of concept , Meta and Google showing AI's revenue impact in advertising; detail addresses spending-sustainability concerns.
Amazon probably did the best job on their earnings call of giving you an idea of what the roadmap looks like going forward, giving you detail, OK, the payback period. We think it's two to three years on these investments that we're doing.
S&P 500 since this was said: ▼ -1.7% · 7,724 → 7,592 · as of 2026-09-10
Bullish
Bloomberg segment on the tech melt-up driving a $3.5T Nasdaq 100 gain in four days; Microsoft noted up 38% from the end of March.
I believe what Microsoft showed in that quarter, I think that's an inflection point for all of tech, because it shows what's happened to monetization and the hyperscowers. It's not just so important for Microsoft, but I think important for the overall sector. We're going to look back and this is going to be a monumental sort of turning point.
S&P 500 since this was said: ▼ -1.7% · 7,724 → 7,592 · as of 2026-09-10
Bullish
Acknowledges 'white knuckle moments' for Microsoft and tech but insists the AI revolution is early.
But the reality is, this earnings season shows you, you're in the third inning. I can maybe even say bottom of the second, you almost go back to where this is all playing out relative to monetization, especially on the hyperscalers.
S&P 500 since this was said: ▼ -1.7% · 7,724 → 7,592 · as of 2026-09-10
Bullish
Asked why Caterpillar 'just blew out numbers'; Ives ties industrials and energy names (e.g., Bloom Energy/Quanta-type plays, per garbled 'room energy, Quana') to the data-center capex build-out.
So this is, remember for every dollar in cap backs, there's a five to six dollar multiply across the rest of tax.
S&P 500 since this was said: ▼ -1.7% · 7,724 → 7,592 · as of 2026-09-10
Bearish
Skyscraper-curse framework vs $725B hyperscaler capex; direct rebuttal of the bear-thesis-is-garbage line. EVENT: will, undated.
so we have seen this cluster of building of data centers And we're going to eventually see a cluster of entrepreneurial errors When these data centers are not profitable
Nasdaq since this was said: ▼ -1.1% · 26,363 → 26,082 · as of 2026-09-10
Bullish
Record S&P close on semis + Hormuz optimism; leadership rotating from chips to adoption winners like Palantir.
So I do not expect that straight line higher in AI stocks as we've seen in the last few years. A much more volatile market but it's still a very structurally strong trend and there will be pockets of gains that we will continue to see here.
S&P 500 since this was said: ▼ -1.7% · 7,724 → 7,592 · as of 2026-09-10
Neutral
Opening answer on whether this earnings season's after-hours whipsaws showed up in the options market, with the Dow coming off a record on the first trading day of August.
Yeah, I will tell you, specifically when you look at an Amazon and when you look at a Microsoft, their one-day earnings realized move was the largest on record. So pulling it back to all the previous quarters, I think it was plus 15% and plus 17%, respectively. But overall, when you look at earnings seasons, specifically as it relates to the hyperscalers, they're beating these implied moves that are happening on options, and that's historically not the case. Typically, options tend to overprice the earnings move, and then they under-realize we're seeing the opposite this time.
S&P 500 since this was said: ▼ -1.9% · 7,737 → 7,592 · as of 2026-09-10
Bullish
Attacked by an American closed-model lab, defended by a Chinese open model.
we managed to defend this cyber attack with an open model coming out of China.
Nasdaq since this was said: ▼ -1.9% · 26,585 → 26,082 · as of 2026-09-10
Bullish
Semis cheap, tech at a market multiple; categorical rejection of the 1999-2000 parallel.
And when you look at the S&P 500 and information technology, it looks nothing like what we had back in 1999 and 2000. So all of a sudden, all this talk about this being a tech bubble is kind of thrown out the window.
Nasdaq since this was said: ▼ -1.9% · 26,585 → 26,082 · as of 2026-09-10
Bearish
Avoid call on AI names priced on speculative cash flows 2-5 years out.
In other sectors or other areas, they're very high in the sense that if you're looking at certain AI names, and you look to the future, some of these companies are trading on speculative cash flows that may or may not materialize two to five years out. So as an investor, you want to stay away from that.
Nasdaq since this was said: ▼ -1.9% · 26,585 → 26,082 · as of 2026-09-10
Bullish
Buildout is the bottleneck; big buyers pay above-average pricing to lock capacity.
And when you look at the bigger companies who need that infrastructure build, they're willing to lock it up and try to lock it up as quickly as they can and pay above average pricing to lock that infrastructure up if they're able to. That's really the bottleneck here is just the build out of the infrastructure and the time that's going to take.
Nasdaq since this was said: ▼ -1.9% · 26,585 → 26,082 · as of 2026-09-10
Bullish
ARK Big Ideas 2026; implicitly bearish incumbent search advertising. EVENT: will, dated ~2031.
We believe AI search can grow from anywhere from 10% today to around 65% of global search traffic over the next five years.
Nasdaq since this was said: ▼ -1.9% · 26,585 → 26,082 · as of 2026-09-10
Bullish
Biggest investment cycle since the late 1800s framing (whisper garbles his name; role stated on tape).
Look, some people say that we are in the biggest investment cycle since the late 1800s. I mean, it is amazing to see sort of the pivot into this incredible digital infrastructure surge around data centers and really the standing up of new industries almost.
Nasdaq since this was said: ▼ -1.9% · 26,585 → 26,082 · as of 2026-09-10
Bearish
Predicts the AI investment boom ends; long-term growth reality vs consensus optimism.
We're going to see the investment boom in AI come to an end and then see how that plays th
Nasdaq since this was said: ▲ +0.6% · 25,914 → 26,082 · as of 2026-09-10
Bearish
US watermark evidence framing; stolen-technology claim vs the China AI rally.
There's a lot of frustration I'm getting from the US side related to the things that China is doing and the stealing that's going on from China from the United States.
Nasdaq since this was said: ▲ +0.6% · 25,914 → 26,082 · as of 2026-09-10
Bearish
~1T of AI datacenter debt issuance is repricing the curve.
He also knows this enormous amount of need for debt in the AI compute data center world where we're talking about potentially another trillion dollars of issuance going on in the market today. The market is finding a home to clear these bonds. It's also raising longer-term rates. It's putting more risk premium into the curve.
Nasdaq since this was said: ▲ +0.6% · 25,914 → 26,082 · as of 2026-09-10
Bearish
First time in his career the largest companies talk about merely staying FCF-positive.
You've taken those companies and you've now made them huge asset gathers, huge asset builders. And they're no longer creating free cash flow. In fact, it's the first time I can remember in the history of an earnings call of some of these largest companies when they're talking about we will stay positive on free cash flow.
Nasdaq since this was said: ▲ +0.6% · 25,914 → 26,082 · as of 2026-09-10
Bearish
The flip side of his AI bullishness: nobody knows if the compute spend earns a return.
But second, I think the huge point is that no one does really no one really knows whether these astronomical amounts of spending on compute Are gonna earn a satisfactory return
Nasdaq since this was said: ▲ +0.6% · 25,914 → 26,082 · as of 2026-09-10
Bearish
AI now dominates equities, IG credit, and 87 percent of venture capital; days after the AI fund blowup.
So the best recommendation today is to be not in AI, because that gets you away from this one factor that is driving markets at the moment.
S&P 500 since this was said: ▼ -0.1% · 7,600 → 7,592 · as of 2026-09-10
Bullish
Microsoft back in AI leadership on retained positive cash flow. EVENT: will (FCF signs).
Yeah, they have. They have. Part of it is that they actually reported positive cashflow and said that they'll retain positive cashflow, where Google and Amazon are already negative and will stay negative.
Nasdaq since this was said: ▲ +0.6% · 25,914 → 26,082 · as of 2026-09-10
Bullish
All three hyperscalers winning in AI; Microsoft the value pick.
Having said that, all three are winners. The only difference right now is valuation. Amazon is the most expensive, Google the second, and Microsoft, the least expensive, all three are winning.
S&P 500 since this was said: ▼ -0.1% · 7,600 → 7,592 · as of 2026-09-10
Bearish
Oracle on the bubble: delayed equity raise, high CDS, falling stock.
On the Oracle side, it's just execution. They need to get data centers built, they're having a problem with that. They need to raise capital, they're having a problem with that.
Nasdaq since this was said: ▲ +0.6% · 25,914 → 26,082 · as of 2026-09-10
Bearish
Owns the host's readback of his written bet: peak AI hit a few months ago, more consolidation and correction ahead, including smaller players. EVENT: will, months-ahead horizon.
Yes, and I'm including the smaller players as well.
Nasdaq since this was said: ▲ +0.6% · 25,914 → 26,082 · as of 2026-09-10
Bearish
Mag 7 protected by cash flows; the shakeout hits the lesser companies (whisper renders Mag as MAX).
So I think when you see a shake out of the AI space, it's not so much the MAX 7 we should be watching, but the lesser companies.
Nasdaq since this was said: ▲ +0.6% · 25,914 → 26,082 · as of 2026-09-10
Bearish
The fast rebound after the AI-fund blowup is FOMO flows, not fundamentals.
It's not like people are rethinking the fundamentals and coming back to investing. They just don't want to be left out of this party.
Nasdaq since this was said: ▲ +0.6% · 25,914 → 26,082 · as of 2026-09-10
Bearish
His Mag-5 marginal ROIC analysis; capex continuing while returns fall.
the drop off in returns on capital is pretty amazing given how big these companies are. So unless they start delivering earnings, commensurate with the tens of billions of investment CapEx, you're going to see a very different kind of company emerging from the mix. More capital intensive, lower return investment capital company.
Nasdaq since this was said: ▲ +0.6% · 25,914 → 26,082 · as of 2026-09-10
Bullish
Calls the chip crash momentum-driven froth, not an AI-fundamentals verdict, at the crash lows.
Well, I think that is the key question here is this correction in the markets, is it driven by fundamentals or is it driven by momentum? And my view is that I think it's driven by momentum
S&P 500 since this was said: ▲ +1.4% · 7,490 → 7,592 · as of 2026-09-10
Bullish
Defends hyperscaler AI capex as sound investment with eventual ROI.
And my view is that it's real that I think there will be a return on all of this cap X.
Nasdaq since this was said: ▲ +2.8% · 25,374 → 26,082 · as of 2026-09-10
Bearish
With 30Y above 5.2 percent he argues risk-free yields beat 50x semis; bonds-over-semis the week chips crashed.
I could now buy a US government bond that pays me 10 percent pre-tax a year. Why the heck would I pay 50 times the earnings for semiconductor stock?
S&P 500 since this was said: ▲ +1.4% · 7,490 → 7,592 · as of 2026-09-10
Bullish
Demand outstripping buildout forces compute prices up, entrenching revenue leaders. EVENT: will.
I think that the price of compute is going to go up and that will provide an advantage to the models that have the most lucrative algorithms that are able to produce the most intelligence per watt
Nasdaq since this was said: ▲ +2.8% · 25,374 → 26,082 · as of 2026-09-10
Bullish
MSFT the adult in the room after Azure 43 percent; relief across the AI trade.
That's why I expect there to be a big sigh of relief across the AI trade. Because if Microsoft is setting the tone as opposed to less responsible actors like Mr. Zuckerberg, or to some extent Google, then maybe we're in better hands than we thought we are.
Nasdaq since this was said: ▲ +2.8% · 25,374 → 26,082 · as of 2026-09-10
Bullish
Pair call: MSFT ~18x and GOOGL ~30x revert toward parity. EVENT: will.
We were expecting a reversion to the mean, and between the Google result and this Microsoft result, we're going to get it.
S&P 500 since this was said: ▲ +1.4% · 7,490 → 7,592 · as of 2026-09-10
Bullish
AI bull who invokes Enron on ~2T of off-balance-sheet SPV debt: the bad ones put the system at risk.
You're going to hear me say that I believe that the investment in AI infrastructure is a wise one, that we're going to get good returns. We're already starting to do that, but I don't like circular financing. I don't like that we're building on leverage.
Nasdaq since this was said: ▲ +2.8% · 25,374 → 26,082 · as of 2026-09-10
Bearish
Flat assertion of Chinese model distillation theft, days before PLTR earnings.
Well, absolutely. I think there's no question since the beginning of the Chinese models that they have benefited through theft, largely what they call distillation attacks. And those distillation attacks have gotten more and more effective over time, more and more automated.
Nasdaq since this was said: ▲ +2.8% · 25,374 → 26,082 · as of 2026-09-10
Bullish
Ottava as category-defining vs rivals stuck in two architectures
“the biggest difference here is our robot is the table. The arms are integrated”
Nasdaq since this was said: ▲ +2.8% · 25,374 → 26,082 · as of 2026-09-10
Bearish
that = robots taking over surgeons' jobs; components someday, not today
“I believe that would be very difficult to accomplish”
Nasdaq since this was said: ▲ +2.8% · 25,374 → 26,082 · as of 2026-09-10
Neutral
AI-related high-tech investment, the economy's most striking feature per the chair
“four quarter growth rates of nearly 20%”
Nasdaq since this was said: ▲ +6.7% · 24,443 → 26,082 · as of 2026-09-10
Bearish
why hyperscaler stocks underperform despite good earnings. backward-looking diagnosis from an index bull, not a sell call
“the free cash flow generation is atro cious”
Nasdaq since this was said: ▲ +4.8% · 24,877 → 26,082 · as of 2026-09-10
Bullish
his read of the de-rating
“It doesn't mean the CapEx cycle is over. It just means we over discounted it”
Nasdaq since this was said: ▲ +4.8% · 24,877 → 26,082 · as of 2026-09-10
Neutral
his window for building AI-native companies. about his business window, not a market call
“opportunity exists for the next 36 months, and then it very well may go away”
Nasdaq since this was said: ▲ +4.8% · 24,877 → 26,082 · as of 2026-09-10
Bearish
reducing his own AI exposure, personal-money admission. PERSONAL-MONEY admission; de-risking, not shorting
“I sold my Google a couple of months ago”
Nasdaq since this was said: ▲ +4.6% · 24,932 → 26,082 · as of 2026-09-10
Bullish
capex continuity across the economy. observed-evidence claim; first name only on tape, surname from the show billing
“we actually see no signs of any slowdown in overall CAPEX”
Nasdaq since this was said: ▲ +4.6% · 24,932 → 26,082 · as of 2026-09-10
Neutral
how the market now treats AI spending
“the word cap X has become toxic”
Nasdaq since this was said: ▲ +4.6% · 24,932 → 26,082 · as of 2026-09-10
Bearish
the historical record on heavy-capex firms. scoped to capex-heavy firms; he stays optimistic on the market overall
“excessive investment lowers returns”
Nasdaq since this was said: ▲ +4.6% · 24,932 → 26,082 · as of 2026-09-10
Bullish
why he doubts capex cuts. 'the cuts' = capex cuts, never rate cuts; keep attached
“I'd be doubtful that there would be the cuts because these companies still have access to the bond market”
Nasdaq since this was said: ▲ +4.6% · 24,932 → 26,082 · as of 2026-09-10
Bullish
where we are in the cycle; nobody slows capex
“It's third inning in the AI revolution”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Bullish
1955 analogy: pull back and you're in Reno. full line ends 'in 1955'; keep the Reno payoff attached or the analogy is opaque
“you're building out the Vegas Strip”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Bearish
tail risk if AI spending reverses. STRENGTH corrected firm->hedged: a diversification warning from an otherwise bullish speaker
“my biggest concern here is too many portfolios are predicated on that spending”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Neutral
Alphabet capex guide he cites. the number is ALPHABET'S guidance, not Payne's estimate
“spend up to $258 billion next year on AI infrastructure”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Bearish
cheap good Chinese models vs US incumbents. journalist analysis; them = OpenAI/Anthropic
“Chinese competitors are proving to be a real risk”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Bearish
Google cash burn on AI spend. hedged with possibly on tape; Tesla named in the same breath
“this is the first quarter possibly since its IPO that we've seen a negative cash flow for the company”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Bearish
AI capex returns. she concedes early signals that AI earnings may be catching up
“Not seeing necessarily a ton of returns on that expenditure just yet”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Bearish
startups switching to cheap Chinese open-weight models
“I can't afford to spend all my tokens on Claude”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Bullish
adoption pace superlative. superlative observation by the CEO of a direct beneficiary
“I've never seen a technology adoption curve like what we're seeing with AI right now”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Bullish
compute supply-demand. same speaker as prior row; name web-verified
“demand for compute will far outstrip supply for the next couple of years”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Neutral
hyperscaler cash math. STANCE corrected bear->neutral: reporter conveying market math, forward projection
“spend more on capex than they generate in free cash flow by next year”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Neutral
hyperscaler capex scale. reported aggregates, deployed in a pro-Apple framing
“spend 800 billion on infrastructure this year, another 1.2 trillion next year”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Bearish
capex into unproven returns. he concedes if we get great returns it will be fine; round rhetoric on the $1T
“a trillion dollars going into data centers”
Nasdaq since this was said: ▲ +4.4% · 24,976 → 26,082 · as of 2026-09-10
Bearish
overbuild risk from less-disciplined competitors. explicitly a monitored tail risk, conditional lead-in on tape; he is otherwise constructive on Alphabet
“this build out goes too fast um, and that lack of discipline ends up getting paid for”
Nasdaq since this was said: ▲ +3.8% · 25,138 → 26,082 · as of 2026-09-10
Bullish
Referring to the early-2025 DeepSeek selloff as the template for the current Kimi K3 scare: cheap Chinese models mean 'more demands will generate more demand for the GPU, for the AI infrastructure, for all the whole sector' @75:30. He also says policymakers on both sides 'don't want the engine to stop now' @74:49.
“So I would say at that moment, it actually provides a very good opportunity to buy into the sector because all that means AI development is pretty good. That will drive small demands instead of slowing the demands.”
Nasdaq since this was said: ▲ +1.5% · 25,691 → 26,082 · as of 2026-09-10
Bullish
He contrasts superstar AI names at 'a TTM PE like 100 times' needing earnings to double or triple, whose robustness 'will be not that safe' @77:41, with AI-related names at 20-30x with 30-40% sustainable growth; expects AI-related sectors' earnings growth above 20%, maybe close to 30% this year @82:32, and says the ripple stays within AI, not consumer sectors.
“The fundamentals is solid earning growth is more visible. And the PE level is only like 20, 30 times. So that name is still cheap.”
Nasdaq since this was said: ▲ +1.5% · 25,691 → 26,082 · as of 2026-09-10
Bullish
He argues the recent AI-trade wobble 'mostly looked like they've wiped out some of the over leveraged, overextended trades from retail investors' @85:48, and with that positioning cleared out the US economy 'still looks very strong' @85:45.
“in a lot of ways we're reasonably set up for further gains in tech shares, for a realisation that the AI boom is going to boost US economic output”
Nasdaq since this was said: ▲ +1.5% · 25,691 → 26,082 · as of 2026-09-10
Neutral
His written note, read by the anchor at 24:03, says 'the outlook for profitability among the hyperscalers is becoming more challenging. They're becoming increasingly asset heavy businesses.' He frames the risk as 'a bit of a loss of patience from investors with regards to hyperscalers recouping some of the spent money' @24:40, while staying constructive: 'As long as the earnings come through, I think those stocks should be able to make progress. But the positioning needs to normalize a bit.' @26:13. The anchor notes he thinks the overall structural AI trend is intact.
“But just to be clear, we're not leaning to the negative side too much. We're just worrying that investors at some point might lose patience.”
Nasdaq since this was said: ▲ +1.5% · 25,691 → 26,082 · as of 2026-09-10
Bearish
Previewing Alphabet, Tesla and IBM results due after the bell. He says the only offset is 'producing super normal growth rates and returns' and warns 'Otherwise, it was a risk for sure that valuations could continue to shrink over the longer term' @41:44. On Alphabet specifically, consensus has cloud growth at '65% year on year' and next-year AI capex 'growing at 3X, what that number was last year', so Alphabet must show accelerating cloud growth to justify the spend @39:15.
“I think if you look at the kind of spending they're doing, then the asset heavy is a concern if they become kind of industrial tech companies, which probably longer term warrants a lower type of valuation multiple.”
Nasdaq since this was said: ▲ +1.5% · 25,691 → 26,082 · as of 2026-09-10
Neutral
On OpenAI's disclosure that two of its models autonomously escaped a sandbox and hacked into Hugging Face @0:00. He continues 'because if it is no longer aligned with human values and it no longer obeys our commands' @2:11, and 'it's not OpenAI that even figured out it escaped... to me, that is the biggest call for regulation' @2:42. A condensed version of this clip also aired in the Jul 22 market-open wrap (harv0722w2_02 @2:27); filed here from the fuller interview.
“And I've been an AI optimist. I think AI is going to increase the number of jobs. I think it's incredibly useful technology. But this is the first thing that just totally scares me”
Nasdaq since this was said: ▲ +1.5% · 25,691 → 26,082 · as of 2026-09-10
Bearish
Syntax garbled by whisper; he prefaces 'I'm not an investor' @7:00. Continues: 'The profits and margins will move from the frontier models if there's more open source and more Chinese things coming' @7:25, and notes Musk also gets Cursor, 'the application' layer @7:19.
“this is an investment kind of hurts these big frontier model companies because they can be undermined by more generic models and people can download the open weights. But it helps Elon Musk because he also has the data centers.”
Nasdaq since this was said: ▲ +1.5% · 25,691 → 26,082 · as of 2026-09-10
Neutral
As Musk's biographer he says Musk 'is totally worried' about AI escaping control @5:34 and 'one of the things he's consistent about is he doesn't like Sam Altman and open AI' @6:43. Host counters that Musk and OpenAI are already 'teaming up at two point five billion dollars a month' @6:50. He praises Dario and Daniela Amodei's constitution approach @4:56.
“I suspect there'll be some partnerships, more partnerships between XAI or SpaceX AI or whatever we call it these days and anthropic,”
Nasdaq since this was said: ▲ +1.5% · 25,691 → 26,082 · as of 2026-09-10
Bullish
Closes the segment: the 'he' is Elon Musk, whom he has just argued benefits from open source, data centers and applications @7:14. The backlash reference echoes the New York moratorium theme running through this week's tapes.
“people need to build data centers in outer space in four or five years because boy, there's a big backlash against data centers these days. So I think he profits from”
Nasdaq since this was said: ▲ +1.5% · 25,691 → 26,082 · as of 2026-09-10
Bullish
Refers to a basket of AI-leaned companies across tech, financials and industrials versus broad S&P margins @4:20. On whether spend continues before returns arrive: 'I think we are going to continue to see that spend' @4:12. Notes funding has shifted from free cash flow to debt and now equity markets, 'a risk that we need to keep an eye on' spotlighted in their 2026 outlook @6:04; metaverse comparison rejected: 'This is not going to happen. This feels very different' @3:31 (referring to a petering-out).
“profit margins in those companies are widening at an even faster pace. So we are actually seeing signs that A.I. investment is beginning to pay off.”
Nasdaq since this was said: ▲ +1.5% · 25,691 → 26,082 · as of 2026-09-10
Bearish
He is underweight the US and has de-risked: 'we've reduced substantially anything that we consider as overvalued' @18:40, rotating into short-term bonds and Swiss equities. Illustrates the AI economics with his own usage: 'I'm paying two hundred dollars a month of subscription with an A.I. company while I'm consuming five thousand dollars a month and tokens' @19:38.
“So this subsidy model is not sustainable unless we see these companies showing real profitability and no need for subsidies anymore done by private equity and VC houses.”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bullish
Asked whether China's cheap Moonshot Kimi K3 model destroys the premium paid for American AI, he says the value sits below the models: 'the iceberg is the chips and the GPUs and the hard drives and the things you put them in and the air conditioning and the cables' @14:31.
“it makes no difference. The models are the tip of the iceberg and they don't really have a very good moat around them”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bullish
Asked whether data-center pushback (New York's proposed one-year build-out ban, 'other states might follow' @1:47) makes the boom a blip. She pegs hyperscaler spend 'expected to be $700 billion this year' @0:34, but also warns that semi supply 'certainly can erode pricing powers and margin' @2:53.
“at the end of the day, we will build and we will still have that massive spend. Again, 700 billion expected this year. It should scratch a trillion dollars of spend in this whole AI build-out next year.”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bullish
Cisco released two free open-weight AI security models (350M and 1B parameters) that day; Patel's other clip: using a frontier model for every attack 'is like saying, I'm gonna take a private jet to go out to a corner store' @1:36. Reporter notes he 'expects those prices to fall... and says that's not necessarily bearish' @2:11.
“Token costs will need to keep going down, but that doesn't mean that the total spending goes down. The spending will go up as the token costs go down.”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bullish
Host: Hut 8 up 20 percent on the week after a second 15-year hyperscaler lease worth nearly $10B, fully commercializing its 1 GW Beacon Point Texas campus @0:00. Genoot: from 'zero contracted AI revenue' a year ago to 'about twenty seven billion dollars of AI contracted revenue, almost one point seven five billion dollars a year' @1:29.
“Power is the key critical resource that allows these this industry to thrive. Our thesis as a company is power fuels technology.”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bullish
Asked about the New York Times report that new data centers will add $6.3B to PJM electricity bills and $29B in utility costs since 2024 @1:49, amid New York's data-center moratorium. He adds developers pay energy bills, fund community infrastructure, and 'if data centers don't pay its full bill, I would tell communities not to bring them in' is the industry standard he claims @3:18.
“It's not true. The key in all of the projects that we're working on is we pay for the transmission upgrades that the utility incurs.”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bullish
Frames 2026 hyperscaler spend revised from $700B to $800B versus $150B in 2023 @2:09, and sizes $7T cumulative as roughly three times Canadian GDP @2:29: 'we think that that will help propel economic growth and earnings' @2:38. Later cites the same buildout as 'a secular tailwind' @5:17.
“The estimate was seven hundred billion. The revised estimate now is eight hundred billion. So very large numbers, significant growth. But here's an eye popping number. Cumulatively through twenty thirty, we expect seven trillion dollars in spending in AI.”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bullish
His firm's machine-learning read of 1,500 company transcripts finds 'adoption rates for A.I. is about 15 percent' with 'lots more room to grow' @5:45. Caveat: companies overspending on AI without deploying it well 'are not going to see the stock returns that investors are hoping for' @6:55.
“So our view is that you'll see outsized returns from those companies that are implementing A.I. both for top line and bottom line management.”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bullish
Touting his committee's just-passed Ratepayer Protection Act (whisper garbles it as 'Rape Prayer Protection Act'), which directs federal resources to help states shield ratepayers @0:50. Calls New York's data-center moratorium 'just isn't helpful' @1:04 and frames the stakes as 'either our values are going to control it, or Chinese' values will @0:33.
“We're about six months ahead of China and AI. People predict. We have a year moratorium, let's say that's nationwide, and we get behind China, we'll never catch up”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bullish
Says the community's school system 'will be the wealthiest school system per capita in the Commonwealth' @3:26 and 'I predict that other county officials in other states are going to be looking at that area' @3:32. Site is an old aluminum smelter taking about a gigawatt of power @4:38; property values in the community will 'almost double' @4:31. Caveat: 'if data centers don't pay its full bill, I would tell communities not to bring them in' @5:17.
“One in my district that's getting a $19 billion 20-year investment for anthropic is going to completely change economically that community.”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bearish
who gets hurt if AI investment sours. STRENGTH corrected firm->leaning: could-destroy conditional; the same sentence denies a broad public bubble
“it could just destroy a lot of VCs and a lot of funds and a lot of PE”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bearish
hyperscalers funding AI with debt
“they're borrowing hundreds of billions of dollars”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bearish
if efficiency improves, capacity strands. explicitly conditional on the efficiency curve; colorful hyperbole
“there's gonna be a lot of data centers that are gonna be turned into pickle ball courts”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bearish
enterprise reality check. scoped to implementation difficulty; on tape he also says it is gonna be amazing
“AI is a lot harder to implement than anybody expected”
Nasdaq since this was said: ▲ +0.9% · 25,837 → 26,082 · as of 2026-09-10
Bullish
He attributes the chip selloff to 'a little profit seeking after the epic run-up that the hardware stocks have had' @0:20, calls Kimi 'a fine model. It's memory intensive and slow, but it is performant' @0:31, and closes that downstream-impact fears on leading models and hardware stocks are 'wildly outblown... Those cases are preposterous' @3:32.
“But this notion that Chinese models are suddenly going to crater demand and change these cost curves, I think it is totally misplaced.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
'I think the frontier models are going to continue to lead as they have for the last many years' @1:08; predicts Chinese open-weights will be shown to be 'trained on distillation attacks on the US models' @1:37, expects 'some level of regulation' against Chinese-model dumping @2:18, and sees leaders able 'to walk the prices down on their models pretty quickly' @3:15.
“But corporate America is not going to chase fractions of pennies per token and embrace Chinese technology for their most critical data assets, market needs, and product needs.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
On Kimi K3: 'the reaction over the last 24 hours is perhaps a little bit premature' @19:28, 'benchmarks are imperfect' @19:42, K3 'is much more expensive on a per token basis than Kimi K2' and 'not particularly token efficient' - 'for a given task, it actually uses many more tokens than an open AI or anthropic model' @20:33.
“And so I think this conclusion that it 's going to lead to price erosion for the frontier is perhaps a bit premature.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
'in the long run, the real threat to the memory bottleneck trade in South Korea comes from China' @50:28; 'That leads you to all the mega listings that are on the horizon, which it now seems the authorities are quite supportive of' @50:45; suggests China 'has been a relatively neglected market, maybe it's worth having that as a hedge alongside the South Korea trade' @53:06.
“And in the second half of the year, we 're going to get a bit of narrative shift in favor of China's AI and semiconductor ecosystem.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
On Kimi K3 as cheap open-weight alternative: 'This is not good news for companies like Azoropic and OpenEye [Anthropic and OpenAI]... they consider they are the king of the world, but now the position is being challenged' @80:38; expects Washington to tighten: 'It will probably make Jensen Huan more difficult to convince... the commerce and the Congress to continue to explore [export] more... advanced models to China' @79:54; hybrid Chinese-basic/US-advanced usage already a trend @80:28.
“It's like a perfect plan B. So that's a clear challenge to the dominance of US AI models.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
On Moonshot's Kimi K3, trained memory-efficient under chip constraints ('proprietary innovations in there to also help with the usage of HBM and GPUs where China is obviously constrained' @0:40); he frames the live debate: does cheap Chinese AI grow the whole market 'or does the Chinese models resource optimization reduce the size of the GPU market?' @2:57. 'showers' = whisper garble of 'shovels'. Duplicate airing exists in the Jul 20 Asia Trade tape; claims filed once, under this clip.
“China has found a software workaround to its hardware constraints and that is what the market is pricing in here on the AI picks and showers trade and that is what will continue to play out.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Neutral
In a soundbite, Detrick notes tech was up ~33% at mid-year (semis and memory up more) and frames the current pullback as a healthy, acceptable unwind rather than a breakdown.
“some conductors up more, memory up more, you know, I mean just letting some air out of that trade is okay.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
Bhardwaj frames AI as a 'realization trade': economists including Fed Chair Warsh say AI is ultimately disinflationary, but she argues the near-term effect is inflationary due to the capex/spending surge, with lots of equity-rotation winners and losers.
“clearly in the near term, it's inflationary first because of all of the capex investment move, all of the spending going in that direction.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
On Kimi K3 and Chinese cheap models displacing frontier models. She tells clients to keep index weight but 'we are definitely recommending taking profits' (@21:05) and, asked about buying a dip, 'only to the index weight? I would say not to go overweight in this area' (@28:14).
“because of the massive underpricing that could then occur That gap that would evolve that could be the apocalyptic moment for many of these Mainstream hyperscaler stocks”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
On Kimi K3 disruption risk: 'to not expect disruption In the the thing that's disrupting the whole entire world would be crazy' (@47:14). Recommends uncorrelated assets (fixed income, film royalties, private credit); calls the semi pullback 'taking a breather is actually a healthy sign' (@49:42).
“So for a while now, we've been diversifying away from some of these hyperscalers”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
Direct answer to Guy Johnson: 'One of the things that's only driving the economy right now is the ai boom... do you think it's going to be worth it?' (@78:14). Adds 'I think ai is going to really change The way we do our work' (@78:39) and that the 'real step function in automation will come with the next generation aircraft' (@79:23).
“Oh, yeah, I do. Uh, ai is critically important to us.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
Replayed Bloomberg Surveillance clip (same morning's programming). Leads with 'It's not as good as ours I don't think' (@46:32); the worry he voices is China flooding markets with cheap compute 'Like they did with trinkets back in the early 2000s' (@46:28).
“They distill it from our models probably which helps on their costs But the reality is they have the electricity grid they have enough chips and they seem to be able to flood their market”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
Concedes 'Some AI created footage that will wind up hitting movie screens' over years and decades (@23:27) but: 'I think there's a strong preference for real film, real creations by humans, by human writers, by human actors' (@25:09) and 'that's what the audience will continue to receive' (@25:28). Also the 'Tom Cruise will always do the stunt' exchange (@24:40).
“we're going to be a very fast adopter of using AI back of house to make our company a stronger company, but very slow as an adopter to use AI on our screens”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
Responding to Kimi K3: open-weight models 'are not necessarily cheaper to run' (@36:26), 'the frontier models are much, much more token efficient' (@36:38), and 'for a lot of tasks, it turns out these frontier models from OpenAI and anthropic are actually just better' (@37:26). Teaser duplicate of this line at 34:25; in-context occurrence filed.
“at the end of the day it's going to be, you know, how efficient is your inference. And I think the frontier labs are still far ahead here. And I think that's really where the value is going to accrue”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
Archival viral clip replayed inside the Taylor segment (original air date not stated on this tape). Fuller clip: enterprise customers 'are livid... I am paying for tokens that create no value. These people are stealing the weights and the alpha of my business' (@41:05-41:13). 'over sailed' = 'oversold' garble.
“And the reason for it is because these models have been completely over, irresponsibly over sailed and the sale is it's dangerous for everyone”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
Asked when 'you wake up one morning and go, okay, there's an overbuilt': 'probably in three to four years' (@4:41). Concedes 'It could' run for quite a while and that tech people say the endpoint is impossible to define; cites fiber precedent.
“So there's likely to be an overbuilt, because every innovation tends to overdo it in the initial phases”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
On the new Chinese K3 open-weight model: 'it turns out the frontier models are much much more token efficient' @1:29, and on US labs keeping the lead: 'i think the u.s labs need for every price performance latency cost they need to have the best model available and actually i think that's more likely than not' @2:46.
“I'll say one thing that I think is a little bit overblown about these open weight models is they're not necessarily cheap cheaper to run”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Neutral
His answer to the enterprise ROI complaints (Khosrowshahi, Karp cited by host @5:23-5:36): 'the market's in kind of a quirky spot where everyone's worrying about paying for tokens' @5:44; at Sierra 'you don't pay per token you pay per outcome you pay per originated loan' @6:32 and 'the solution to a lot of the problems with ai tokenomics is we don't have a mature applied ai market' @7:52.
“i believe where the world is going is paying for outcomes”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
Framed with 'it's early days, Maria, but it's very, very exciting' @0:18 and the reversal of history: 'Historically, healthcare has been somewhat slow to adopt new technologies. And we are seeing the exact opposite with AI' @0:22; examples include a study finding pancreatic cancer signs 'up to 16 months ahead of human readers' @2:15.
“There's some recent research that suggests that we're adopting this technology about two times as fast as any other industry in the American economy.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Neutral
A live on-air correction of the host's assertion that 'you're already seeing the longevity of people, a person's lifespan get impacted' @5:45; he had already cautioned 'There is very limited evidence, and when I say evidence, I mean robust peer-reviewed evidence... of the impact of AI in real-world patient care' @4:21, while adding 'We're all optimistic that that's what's going to happen' @6:05.
“So, it's, again, too early to say. I think that's too strong a claim to be supported by existing evidence that AI is actually prolonging folks' lifespans today.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
Asked which sectors can outperform, he answers 'The easiest one to talk about is still technology' @1:54 and describes AI as 'a five-layer cake' (models, application, memory, energy) with hyperscaler capex 'feeding into other sectors' including energy and industrials @2:10.
“And right now, the AI companies or the ecosystem related to AI is really driving earnings from my perspective.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
Reacting to OpenAI's brutal week (Apple lawsuit, Atlas browser shutdown, Fiji Simo exit, Oracle downgrade citing OpenAI). He argues bubbles pop in order: apps disappoint, infrastructure overbuilds, capital markets close: it is massively overspent here. And these valuations are gonna come down dramatically @10:57; the real risk isn't that OpenAI loses, is that nobody wins enough to justify these valuations @10:35. He still calls OpenAI and Anthropic amazing companies even down 80 percent @28:13.
“This does feel like Q4, 99. And then March of 2000 is now in plain sight.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
He calls it AI dumping from China @17:51: free Chinese models went from less than a third of all traffic in late 2025 to about two thirds recently, subsidized, on cheaper power and chips. Ed supports with pricing data @25:15: GPT 5.6 at $45 per million output tokens vs DeepSeek at 87 cents. Scott still thinks US premium frontier models can compete at the high end (@27:19).
“What the Chinese automobile industry did to Detroit in about three decades, it feels like Beijing is doing to USA AI frontier models in about three months.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
Scenario math combining hardware shutdown, an 80 percent price cut, and eMarketer's 90 percent ad-revenue miss. He does not predict bankruptcy: I think the more likely scenario is that there will have to be some sort of massive restructuring, some sort of M&A event. Probably it gets sold to a larger AI company @37:43. Biggest worry: if I were an investor in OpenAI, it would be the China problem. It would just be the pricing @38:26.
“they have said that they're gonna be profitable by 2030. But when we put all these estimates together, what we find is that actually, no, they won't be cashflow positive. They will be losing about $165 billion a year.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Neutral
Clip begins This AI thing better work out ... Because if that doesn't work out, then your portfolio will be in trouble @47:50, and ends there's one factor staring all of us right in our eyes. And that is AI is literally everywhere @48:17. Ed also cites Slok's stat that AI accounts for nearly half of all investment grade bond issuance @47:13.
“That's why ironically, the best investment recommendations today is the new 60/40 is really to do 60 maybe AI and 40 non-AI.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
Asked about New York blocking data centers; the ensuing back-and-forth with the host includes 'We need data centers in this country' @4:50 and 'New York doesn't want them. The jobs will go somewhere else' @4:58, though turn boundaries in that stretch are blurry in the transcript, so the quote is restricted to Moore's clearly cued answer.
“Yeah, ridiculous. It's like when they stopped fracking. They didn't get the oil.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
Asked whether the Kimi open-source model release changed the neocloud story, he invokes Jevons Paradox: cheaper models 'just create more demand for AI' @0:54 and 'The demand for compute is sky high' @1:39, adding his wife just adopting ChatGPT means 'the masses are coming' @1:52. Whisper garbles neocloud as 'Neil Clout', IREN as 'iron', Cipher as 'Cypher', Kimi as 'Kimmy'.
“So nothing has really changed from my kind of long-term bullish take. I still own iron, I still own Cypher. Probably my favorite Neil Clout is Hut 8.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bullish
He opens 'I think there's plenty to worry about' @0:37 and expects consolidation as in the early-2000s tech boom @1:01, but says 'in every earning season over the next couple of quarters, the demand for AI is just continuing without slowing down' @1:18.
“However, we continue to believe we're in the early innings of this, and I think the demand for compute and what's happening with chip makers and semiconductors, this is just the beginning.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Neutral
Asked if some names like Oracle are more entangled in the circularity problem: 'Absolutely, without question' @1:52, some names carrying more of it than others; unlike 2025 when anything AI-adjacent rallied, 'as we continue to go throughout the year, you're going to see more winners and losers' @2:22, urging close attention to guidance and margins.
“it's no longer an AI trade where a rising tide is simply lifting all boats.”
Nasdaq since this was said: ▲ +2.2% · 25,508 → 26,082 · as of 2026-09-10
Bearish
Companies not getting the ROI they anticipated.
“a lot of CFOs are gonna say play time is over. I need you to either justify the return on investment in our token consumption or to reduce it.”
Scott Galloway · NYU Stern / Prof G Media · tech-ai @AI spendNasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
Rebuts the 'top-3 companies never go bankrupt' point: drawdowns are the risk.
“Amazon lost 92% of its value, so did Cisco. They didn't go bankrupt, but they lost 92.”
Scott Galloway · NYU Stern / Prof G Media · tech-ai @AI mega-capsNasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
Front-end players admit they overestimated demand generation.
“the only people creating demand right now are entropic and open AI”
Scott Galloway · NYU Stern / Prof G Media · tech-ai @AI demandNasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Neutral
Says US labs still lead, but Moonshot's K3 is near-frontier and cheap.
“the Chinese AI companies are catching up with the frontier US labs very quickly now.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Neutral
Sacks endorses Demis Hassabis's FINRA-style SRO proposal over a government AI agency, laying out five conditions (broad industry representation incl. open source, frontier-only, catastrophic risk only, voluntary-first, must substitute not add). Warns the SRO is 'the opening bid' unless kept 'pure.'
“if my choices are between FAA for AI, or what I would call the DMV for AI, I would much rather go for Demis' SRO for AI, the self regulatory approach”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
Sacks reiterates his Oct 2024 claim, citing a Politico piece ('inside anthropic state by state plan to ratchet up AI rules') and Anthropic funding groups like Public First to argue the company deliberately drives a patchwork of ever-stricter state AI rules to pull the ladder up.
“I tweeted that anthropic is running a sophisticated regulatory capture strategy based on fear mongering.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bullish
Rebutting NY Gov. Hochul's statewide hyperscale-data-center moratorium, Sacks calls her claims false, defends behind-the-meter power, land-use efficiency, closed-loop water ('same water as two and a half In-N-Out burgers'), and natural-gas cleanliness.
“data centers are honestly one of the best things we could be building as a nation.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Neutral
Chamath backs the SRO because 'a torrent of money' will try to buy regulatory capture on both political sides; establishing self-set rules and superseding federal oversight avoids a duopoly and killing open source.
“I think it's really important. And I hope it happens quickly.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
Building on Ramp CEO Eric Glyman's clip that token spend among Ramp customers grew 21x in a year, Chamath calls uncontrolled AI token spend 'a money burning furnace' that will eventually surprise CFOs.
“if things are 21Xing every few months, somebody's going to miss a quarter. I don't know who, but somebody.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bullish
Chamath says the US is 'massively short electrons,' citing a PJM auction that sought 7-8 GW but drew only ~156 MW, and Elon's behind-the-meter turbine workaround; he frames power as the binding constraint on AI.
“By 2050, the United States of America will be 2.5 California's worth of energy in deficit.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bullish
Ives says if he thought AI were in the 'seventh, eighth inning' he wouldn't have launched his firm; he thinks investors still underestimate 'the scale and scope of this broad theme.'
“It speaks to my view, third inning, one out, man on first.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bullish
Frames the selloff as an 'A.I. mid-cycle slowdown' second-derivative repricing plus a genuine deleveraging (Korean margin closures, quant/multi-strat losses, momentum-factor unwind): 'We've done enough based on what I think is there. And I think we'll probably start to form some sort of a bottom here.' @4:39. Micron cited as 100 to 1200 and back to 800 that morning.
“I think somewhere between this 30 and 60 percent retracement, they're going to hang in there and they'll start to move higher again. But I think the fireworks show is over.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bullish
Says a very large share of his portfolio is in physical AI and robotics (@46:17); argues the software side of AI is over-crowded and picked-over. Visser concurs, playing robotics via the memory/semis complex.
“the next big kind of move of capital within the AI trade will be that robotics”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
Slok frames the current tech correction as a repricing of AI ROI timing: 'markets are now saying, show me the money, show me the returns' @1:05, and notes S&P 493 profit margins 'have literally done nothing for the last 10 years' @1:17, so the inflection must broaden beyond the Mag 7 and hyperscalers.
“that the correction we're seeing at the moment exactly is a reflection that maybe the returns on the ROI, on AI investments, are going to come, but they're just going to come slower than what markets are pricing.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bullish
On the day MoonShot's Kimi K3 'is sinking AI and semiconductor stocks' (host @0:00), Bloxham says it is 'going to, I think, create pressure for those big U.S. model developers like anthropic open AI and Google' @1:06, while noting it's 'still early days' on how good the model is @0:25.
“corporates are getting increasingly sensitive and focused around the cost of using AI models. So anything that can potentially deliver similar or equivalent performance for substantially less money is going to become more popular.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bullish
'There's a captive audience in global south countries, many of which are economic partners to the Belt and Road Initiative' @14:50; he ties it to 'a parallel discussion about overvaluation on the future financials of US AI giants, that could be a story that plays out over the next few years' @16:33-16:42.
“because if China can compete on cost grounds with the US, with Japan, with others, that's really an open road, I think, for China to make some serious inroads”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
Asked if the 5-6 week selloff is any closer to the bottom: 'I think probably not' @19:35. TSMC and ASML 'report excellent earnings... that didn't help because it didn't assuage those cost concerns' @22:07-22:19; 'just more of the old big earnings by themselves, that's not enough' @23:28. Nikkei down ~4.5% on the day per host @20:43.
“Meanwhile, they also still seem so elevated on a historical basis that we 're some way away from capitulation.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bullish
With the Nasdaq indicated down ~480 points and peak momentum stocks 'coming down about 20% over the last four weeks' @1:47, he defends the ~$3 trillion three-year AI capex estimates as reasonable 'because there is tremendous demand for the end product. Chat GPT is at 1 billion users' @1:21, demand for Claude in enterprises 'almost going vertical' @1:34; the alpha is in 'rotating bottlenecks' - cooling equipment, backend equipment, process controls @2:01.
“Strong conviction on the AI theme overall, especially for the next six to 12 months.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bullish
Attributes the selloff to positioning and valuation, not a broken thesis.
“it is fundamentals that are still holding up and it's the valuations and the nervousness around positioning is what is driving this sort of volatility.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
Cheaper second-mover models threaten AI valuations.
“this really underscores the fundamental risk to the AI trade, the second mover advantage. If we can do the same or more with less than what does it mean for AI stocks.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
Book disclosed: frames the shift as ideal for AlphaSense; enterprises now scrutinize token ROI.
“this era of AI at all costs is over. And in fact, enterprises are demanding evidence that their AI spend is actually producing returns.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Neutral
Moonshot Kimi's disclosed capabilities match or exceed recent US models; still not ahead.
“China has not surpassed the US, but it is certainly competing at a closer neck than we might have thought.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
On $700B of AI investment: AI 'is distorting the markets much more than it is the C suite' (@0:35); 'Every time this happened before the markets will get to the right place. There will be dislocation' (@1:01). He does not call timing: 'where the dislocations... I don't know'.
“At some point investors will demand a return. And if you're worth a trillion dollars and you traded 30 times that's thirty five billion dollars in income you need”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Neutral
Record hyperscaler debt and equity issuance; crowding out other corporates 'Not yet' - 'there's a ton of liquidity. I am surprised that investors particularly on the debt side have been over describing [oversubscribing] by multiples' (@5:26-5:37, garbled).
“As soon as there is a turn in the hyperscalers I think we'll have a very different conversation”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
Hyperscaler capex already priced to go from ~$450B to ~$725B ('maybe that'll break a trillion next year that's already priced in we need to beat those expectations'); until Meta/Google/Microsoft guide higher, 'the market is going to be skittish because SpaceX kind of burned them' @11:23. IBM's spending prioritization and Meta's compute-sale reports read as possible early capex warnings.
“as soon as we see capex roll over I agree that it is game over”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bullish
Semis/hardware down with the momentum factor in first-half July and the day of China's K3 open-source model release; he says 'our base cases that probably last for another couple of weeks or so' @0:59 before the zoom-out call, and 'you never know if this is like the bottom' @0:33.
“But when you zoom out, yeah, I think this will be perceived as a good buying opportunity.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
Conditioned on his team benchmarking K3: 'we need to test this model and run it through our own internal benchmarking' @1:59; he later softens further, 'even with the labs, I don't know that it's a disaster yet for them until we test the model' @4:33.
“But if the reports are correct, then it's really negative for the frontier labs for the frontier models here in America.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bullish
Asked whether AI spending is inflationary he answers 'No, I do not' @2:21, sizing data-center construction at '80 billion to 100 billion dollars' against '$2.2 trillion' total construction @2:35, and concludes 'I don't think by itself that represents an inflation risk' @3:01.
“I think there's more than enough room for the US economy to accommodate and increase in spending on AI and increase in spending on people that are gonna be hired to work at developing AI programs and the like.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Bearish
Immediately preceded by 'But the future for AI is very, very bright. Unfortunately for companies, that's going to invite more competition' @4:03 - the bright macro future is exactly what invites the margin pressure.
“That would put pressure on margins and it might well be that some of the returns that we expect to see from AI on a company by company basis come in lower than anticipated.”
Nasdaq since this was said: ▲ +2.2% · 25,520 → 26,082 · as of 2026-09-10
Neutral
Opens: 'I'm very focused on AI for public good... this is not hostility to the technology' (@0:00). Aired as the foil for Payne's monologue attacking the moratorium.
“but it is simply saying, we want to take a pause. We want to step back and make sure that New York is not only just first, but we're the first to get it right.”
Nasdaq since this was said: ▲ +0.8% · 25,882 → 26,082 · as of 2026-09-10
Bullish
Monologue argues 'there is a campaign underway to derail U.S. efforts in artificial intelligence... and it's political' (@0:36); cites NY's 9 data centers at 30 cents/kWh vs Texas 125 at 17 cents (@1:37) and TSMC's added $100 billion Arizona investment 'last night' (@2:10).
“but we have no time to waste. We must win the race if we want to remain with our peace and our prosperity.”
Nasdaq since this was said: ▲ +0.8% · 25,882 → 26,082 · as of 2026-09-10
Bullish
'infrastructure is where we're leaning into because clients have been under allocated' @31:14-31:21; 'Our family office survey, less than only 1% of family office asset allocation is catered towards infrastructure' @31:22, citing AI, electrification and 'reindustrialization of society' @31:44 as the secular backdrop, plus power and transmission bottlenecks @27:29-27:52.
“We think that's under allocated primarily because infrastructure gives you inflation resilient income backed by multi-year cash flows that are backed by secular themes related to the artificial intelligence rollout,”
Nasdaq since this was said: ▲ +0.8% · 25,882 → 26,082 · as of 2026-09-10
Bearish
Asked if New York's first-in-nation data-center moratorium is 'the first of many': 'Yes. Now, I don't think it's sort of an absolute yes' @12:59. He notes AI regulation 'is really not that partisan... it's red states and it's blue states' @18:53, seven in ten Americans oppose local data centers, and ~$130B of projects already delayed or blocked in Q1 2026 (host framing). He does see a resolution path if operators bring their own power, closed-loop water, and community benefits @14:30-16:19.
“when you look at the unpopularity of AI across the board, then of course, politicians are going to pass bills and regulations and sign executive orders that come off as anti-AI”
Nasdaq since this was said: ▲ +0.8% · 25,882 → 26,082 · as of 2026-09-10
Bullish
Data centers use 'About 5% of all the electricity in the United States' @0:53 but spread grid costs; he blames price rises on offshore wind and nuclear closures @1:17-1:26, and says data centers 'probably accounted for around half of all economic growth' in the early months of Trump's term @2:39. In Loudoun County they pay 'almost half of all taxes' @5:30.
“The power they use and they buy, that pays for the electric lines, that pays for generators. And study after study proves this is reducing prices for the average human in America.”
Nasdaq since this was said: ▲ +0.8% · 25,882 → 26,082 · as of 2026-09-10
Bearish
Direct answer to 'Do you buy that?' after the Hochul clip claiming companies 'will come back to New York on our terms.' He contrasts with Loudoun County, Virginia, 'the global epicenter of data centers' @5:14, and says the animus 'is actually just general animus against the tech industry... misplaced on these real concrete things that are bringing real benefits to local communities' @5:52.
“No, I mean, I think businesses are pretty smart. And when a whole state tells them you're not welcome here, I think some of those business are gonna have second thoughts next time around”
Nasdaq since this was said: ▲ +0.8% · 25,882 → 26,082 · as of 2026-09-10
Bullish
Played clip (original venue/date not stated on tape) defending her first-in-nation statewide data-center moratorium: 'So they can go there right now. I think when they look for the state that has the most clear road ahead that we remove uncertainty... you'll be able to move forward' @4:34-4:49. Glock immediately disputes it on the same tape.
“they'll come back to New York. I have no doubt about it. But they'll come back to New York on our terms.”
Nasdaq since this was said: ▲ +0.8% · 25,882 → 26,082 · as of 2026-09-10
Bearish
Very bullish copper long-term (structural deficits) but $6 is not low; to the degree copper pricing is AI-narrative-driven (data centers), a reversal is his buying opportunity
“if the AI story goes into reverse”
Copper since this was said: ▲ +3.8% · 6.29 → 6.54 · as of 2026-09-11
Bullish
'I actually get frightened when I see states say we're going to do a moratorium. That's not the answer.' (@5:13, the direct answer to New York's data-center freeze, POS-340); a trillion of capex from six companies 'not enough'; contracted ~1GW of power in Pennsylvania last quarter; $50-60B per gigawatt; China building ~100GW nuclear and ~100GW solar; wants a compute futures market
“My worry about it is not a bubble. We don't have the ability to build fast enough.”
Nasdaq since this was said: ▼ -0.7% · 26,269 → 26,082 · as of 2026-09-10
Neutral
IBM vice chairman the morning after the pre-announcement (-$69B day); quantum-rally context; Arvind called the pause 'relatively temporary'; principal defending his own company (disclosed)
“our software is not being disrupted by AI”
Nasdaq since this was said: ▼ -0.7% · 26,269 → 26,082 · as of 2026-09-10
Bearish
'people are spending exorbitantly more actual dollars on tokens than they would have thought they would be spending at this time' (@1:24); 'people are starting to think about what's been the return on investment in my AI tokens? Are they a good investment?' (@1:56); construction labor crowding-out too
“If your budgets X and you spent point nine X on tokens”
Nasdaq since this was said: ▼ -0.7% · 26,269 → 26,082 · as of 2026-09-10
Neutral
Zoom CFO on Bloomberg Brief; Zoom's early Anthropic stake now worth ~$1B (on tape); manages cost with multiple model providers plus its own SLM
“there's not a tech CFO alive that isn't very focused on on tokens”
Nasdaq since this was said: ▼ -0.7% · 26,269 → 26,082 · as of 2026-09-10
Bearish
Asked about New York's statewide data-center ban: notes '69 local jurisdictions throughout the country that have active moratoriums against data center development' (@26:58) plus shrinking-footprint obsolescence risk - 'they'll probably get smaller... you don't need as much space, and there's also the risk of it going to space' (@27:20). Also flags AI as 'a risk for the office sector' (@26:23).
“So we do not have a data center in our portfolio, and we're not actually looking to grow into the sector. And the reason for it is because it's really difficult to underwrite the regulatory risk, as well as the obsolescence risk that exists.”
Nasdaq since this was said: ▼ -0.7% · 26,269 → 26,082 · as of 2026-09-10
Bearish
His oil analogy for frontier token pricing: OpenAI also $26, Anthropic's latest $56, Elon $1, Zuck $1.50, Demis and Sundar $1, 'The Chinese will sell it to you for $0.50. So this rationalization has to happen.' (@2:10); explicitly declines the host's invite to call an LLM break; discloses his seat: 'My entire job is selling enterprise opportunities.' (@0:24)
“Let's say a barrel of intelligence, which is a million tokens. You can buy that barrel for $26 from Anthropic's really good model.”
Nasdaq since this was said: ▼ -0.1% · 26,107 → 26,082 · as of 2026-09-10
Bearish
'The CEOs and the CFOs, in my opinion, probably have no idea how much token maxing is going on inside of their organizations.' (@3:20); 'if you see a bunch of public companies in the next few quarters miss because of an OpEx miss' the market discounts that technology (@5:07); said the day IBM's prelim revenue wiped out $69B, the most since at least 1968 (Bloomberg tape @14:24)
“I suspect what will happen is one day you're going to have a miss, and EPS will be off by a few pennies. And the CEO will say to the CFO, what happened?”
Nasdaq since this was said: ▼ -0.1% · 26,107 → 26,082 · as of 2026-09-10
Bullish
First congressional testimony as Fed chair, on the AI buildout the day New York froze new data centers; adds 'We at the Fed are monitoring the implications for employment and for inflation.' (@25:43); clip aired within Bloomberg Balance of Power 07/14
“Listen, this seems inevitable. That which we're now calling A.I. investment will soon just be called investment.”
Nasdaq since this was said: ▼ -0.1% · 26,107 → 26,082 · as of 2026-09-10
Neutral
First-in-the-nation statewide one-year moratorium on new large-scale AI data centers, discussed on Bloomberg's Odd Lots (clip aired on Balance of Power); PSC framework 'done within this year of the moratorium so we can have certainty for businesses' (@24:25)
“But you come here to bring your own power source or you have to pay a premium to use our grid.”
Nasdaq since this was said: ▼ -0.1% · 26,107 → 26,082 · as of 2026-09-10
Bearish
CNBC contributor, discloses owning IBM on air ('I own it, so I'm not happy'); software revenue +5% vs double-digit hopes
“it's coming to pass that AI is having an impact on software”
Nasdaq since this was said: ▼ -0.1% · 26,107 → 26,082 · as of 2026-09-10
Bearish
On Apple's trade-secrets suit against OpenAI over its hardware business. He says Apple has OpenAI 'basically dead to rights' and that an adverse ruling could unwind the hardware effort: 'if that has to roll back from the beginning, that's going to be pretty damaging to its IPO case' @6:32. Also: 'one by one, it seems to be alienating its partners' @7:10. Teaser duplicate of this line at 0:00-0:33; in-context occurrence filed.
“I expect this to be a fierce battle. I don't think Apple is interested in settling. I think they'll try to take it all the way through to get to a judge ruling”
Nasdaq since this was said: ▼ -0.1% · 26,107 → 26,082 · as of 2026-09-10
Bearish
X post relayed by CNBC's Kate Rooney (conduit): accuses Musk of overpromising space datacenters to public investors
“Homeboy, you're the one selling public market investors on short-term space data centers.”
Nasdaq since this was said: ▲ +0.8% · 25,873 → 26,082 · as of 2026-09-10
Bearish
Commodities lens on AI: falling compute price signals possible overbuild
“AI compute, it's a cyclical commodity. And the point I say too, is you look at the price of AI compute, it's actually going down now, which is an indication you may have overbuilt already, but it's not going high.”
Nasdaq since this was said: ▲ +0.8% · 25,873 → 26,082 · as of 2026-09-10
Neutral
On why the straight-up move in AI names has stopped; 'I think that's a good thing'
“One is it's more expensive than we thought. And two, not everybody's gonna win. So the venture capitalist mindset is starting to set in”
Nasdaq since this was said: ▲ +0.8% · 25,873 → 26,082 · as of 2026-09-10
Bearish
His own query on S&P 493 EPS lift from AI since 2024
“the answer as far as all publicly available data was that the actual ROI was somewhere between zero and 2%.”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bearish
On corporate AI token spend
“At some point, you're going to have to show an ROI that's above the risk free rate of return. Otherwise, you're going to have some angry investors on your hands.”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bullish
Direct rebuttal to Chamath's zero-to-2% ROI math
“I think there probably isn't direct ROI, Chamath, to your point, but I think we're so early, nobody cares.”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bearish
After describing his own token costs dropping 95% on open-source models
“everybody's going to learn how to drop the price by 95%.”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bullish
On how long the AI trade has left
“There's probably a bubble in the people calling it a bubble. But like we don't even have the fleets of robots out there. And we think that this is going to keep going for a very long time.”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bearish
45 years managing money; turned defensive at record highs
“I always get concerned when everyone's in, everyone loves the story, valuations are at all-time highs, and you're really running an economy that, in a sense, is a one-legged stool, one very big fat leg, which is as AI spend”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bearish
The 2000 telecom-capex analogy for hyperscaler spending
“that's what happened in 2000, Greg, and I mean, that's when we saw the end of the tech run there is once all that telecom spending slowed down and everybody looked around and said”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bearish
“The reality though, it is a capital investment cycle. And eventually, many of these big hyperscalers are going to decelerate that investment spend. When that comes to fruition, many companies will their multiples will start to contract.”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bullish
The other half of his split call
“so I fundamentally think that AI as a sector and therefore the demand for the semiconductors, the data center businesses, all these things that people worry about, I don't think that's a bubble. I think that's for real.”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bearish
Calls the current move an unwinding of the AI bubble
“The AI stocks I think are just grossly overvalued based on future earnings, but may or may not come.”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bullish
He first called the mid-cycle slowdown in late May; now calling its end
“So I'm calling the AI mid-cycle slowdown is over”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bullish
Invites being held accountable on tape: insatiable demand, not enough supply
“This is why there is no solution to compute. If you think there is excess capacity of compute, you are absolutely wrong.”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bullish
Host on the record while asking about short-seller critiques
“You're bullish on this. I'm bullish on this. Many people are buying these names.”
Nasdaq since this was said: ▼ -0.8% · 26,282 → 26,082 · as of 2026-09-10
Bearish
The distinction he draws:
“But I'm a bear around AI stocks right now. I'm bullish on AI technology.”
Nasdaq since this was said: ▼ -0.5% · 26,207 → 26,082 · as of 2026-09-10
Bearish
Jeff Currie says the hyperscalers have become cyclical old-economy industries spending at the same rate miners and oil producers did at the 2014 peak, implying the same over-investment risk.
the hyperscalers they are now cyclical industries like commodities their old economy now and they're spending just at the same rate all those miners and all those oil guys did back in 2014
Nasdaq since this was said: ▼ -0.5% · 26,207 → 26,082 · as of 2026-09-10
Bullish
“I think we are in the early stages of the AI boom”
Nasdaq since this was said: ▼ -0.2% · 26,121 → 26,082 · as of 2026-09-10
Bearish
“every single one of these vertical apps expanded into categories that was previously served by companies building on top of Anthropic's own models”
Nasdaq since this was said: ▼ -0.2% · 26,121 → 26,082 · as of 2026-09-10
Bearish
“these enterprises are at risk of transferring their knowledge, their know-how, their trade secrets, their customer data to these model providers who might eventually decide to compete with them”
Nasdaq since this was said: ▼ -0.2% · 26,121 → 26,082 · as of 2026-09-10
Bullish
Goldman's 12-month dollar-yen forecast:
“if you wrap the open source model with our software factory, it was 16.4 x cheaper”
Nasdaq since this was said: ▼ -0.2% · 26,121 → 26,082 · as of 2026-09-10
Bearish
“by handing it over to a model company to then combine with other people's data, you are effectively commoditizing the asset”
Nasdaq since this was said: ▼ -0.2% · 26,121 → 26,082 · as of 2026-09-10
Neutral
“the implications of AI for the business models of the Magnificent Seven, in particular Microsoft, Alphabet, Amazon, and Meta, far from clear, could be great for them. Could not be so great, right?”
Nasdaq since this was said: ▼ -0.2% · 26,121 → 26,082 · as of 2026-09-10
Bearish
“The way to describe this historically is this is the fiber overbuild of 1999.”
Nasdaq since this was said: ▼ -0.2% · 26,121 → 26,082 · as of 2026-09-10
Bearish
“this is a growing bubble that is nearing a point of maybe not collapse, but certainly massive course correction.”
Nasdaq since this was said: ▼ -0.2% · 26,121 → 26,082 · as of 2026-09-10
Bearish
On AI-lab finances:
“When basically open AI is seriously in trouble. And in my view, anthropic is not too far behind.”
Nasdaq since this was said: ▼ -0.2% · 26,121 → 26,082 · as of 2026-09-10
Bearish
Asked about the Iran war's likely duration (Jul 2):
“the result of stronger returns in the second quarter, I think has led to some pretty serious crowding within the AI trade”
Nasdaq since this was said: ▼ -0.2% · 26,121 → 26,082 · as of 2026-09-10
Bearish
Market shakeout thesis: "the market is ruthless and the market is opportunistic" (@3:09); AI companies "remind me of Amazon" - long-term bright, shakeout first
Nasdaq since this was said: ▲ +1.0% · 25,833 → 26,082 · as of 2026-09-10
Bullish
Dot-com comparison, on-set board 'A Boom Without the Bubble': then P/E +300% vs down now; prices +780% then vs +160% now (“some stocks up 1,000%”)
S&P 500 since this was said: ▲ +1.4% · 7,483 → 7,592 · as of 2026-09-10
Bearish
“Once the AI bubble has burst, I'm telling you the US economy is going to hit a break and go into the recession… even a little tiny hint that the capex story is over… [the NASDAQ] will be down 20% like in no time”
Nasdaq since this was said: ▲ +1.0% · 25,833 → 26,082 · as of 2026-09-10
Neutral
“Are we really going to outsource the battlefield of this country to the consensus view in Silicon Valley? That is effing insane” - on model weights and the Department of War
Nasdaq since this was said: ▲ +1.0% · 25,833 → 26,082 · as of 2026-09-10
Bullish
“I think that the market is ruthless and the market is opportunistic and the market sees in a case like Anthropic and OpenAI it sees the heavy spend and it sees the near-term losses but it sees the long-term very bright very rosy outlook. The AI companies remind me of Amazon”
Nasdaq since this was said: ▲ +1.0% · 25,833 → 26,082 · as of 2026-09-10
Bearish
“I mean, I don't know the answer. We probably are in a bubble”
Nasdaq since this was said: ▲ +1.0% · 25,833 → 26,082 · as of 2026-09-10
Bullish
Lee says dexterous robots will transform residential construction within a couple of years.
But I think in a couple of years, like residential construction will be transformed. Like I think homes will be built by artists and robots
Nasdaq since this was said: ▲ +1.0% · 25,833 → 26,082 · as of 2026-09-10
Bullish
Lee says semiconductors may be in a structurally new cycle driven by robots, with good visibility for 2026-27.
I might guess that semiconductors are in a new cycle, a new story
Nasdaq since this was said: ▲ +1.0% · 25,833 → 26,082 · as of 2026-09-10
Bearish
“a lot of these earnings that we're seeing that they hinge on these contracts with these AI companies whose ability to actually pay out on those contracts”
Nasdaq since this was said: ▲ +1.0% · 25,833 → 26,082 · as of 2026-09-10
Bearish
Railroad money 1873 ≈ AI data-center money today (“about the same”); “500 new railroads were created and by the end of 1872 only a hundred were able to pay dividends”; AI “not going to make money in a straight line”
Nasdaq since this was said: ▲ +1.0% · 25,820 → 26,082 · as of 2026-09-10
Bearish
On Chinese exports and Europe's industrial base:
“the amount of money that was going into railroad construction is about the same as the amount that's going into the construction of data centers”
Nasdaq since this was said: ▲ +1.0% · 25,820 → 26,082 · as of 2026-09-10
Bearish
Elson tells Isaac that Zuckerberg is spending heavily on AI data centers and is the worst of the hyperscalers at communicating how it will earn a return.
he's doing probably the worst job of all the hyperscalers in communicating
Nasdaq since this was said: ▲ +2.9% · 25,359 → 26,082 · as of 2026-09-10
Bullish
Far more demand for intelligence than supply; AI trade huge
Ankur Crawford · Alger · Bloomberg (Alger)Nasdaq since this was said: ▼ -1.6% · 26,518 → 26,082 · as of 2026-09-10
Bullish
Lean into AI-infrastructure recipients (power, data centers)
Kevin Mahn · Hennion & Walsh · Fox / Hennion & WalshNasdaq since this was said: ▼ -1.6% · 26,518 → 26,082 · as of 2026-09-10
Bullish
Own the AI power-generation recipients, not just Mag 7
Stuart Kaiser · Citi · Fox / CitiNasdaq since this was said: ▼ -1.6% · 26,518 → 26,082 · as of 2026-09-10
Bullish
AI chips supply-limited for years; demand enormous
Christophe Fouquet · ASML · Bloomberg (ASML CEO)Nasdaq since this was said: ▼ -1.6% · 26,518 → 26,082 · as of 2026-09-10
Bearish
Visser says semis, cooling and AI-infrastructure names face three to six months of mixed news after a hoarding-driven spike.
And now we're going to be in a phase for the next three to six months where a lot of these semiconductor names and a lot of the cooling names and a lot of all the things on the infrastructure, we're probably going to have equal news of disappointment with equal news of surprises
Nasdaq since this was said: ▲ +0.7% · 25,889 → 26,082 · as of 2026-09-10
Neutral
Visser says the cost of producing tokens is about 4x a year ago due to compute bottlenecks.
The other side is the cost of producing the tokens is going up dramatically. And when I mean dramatically, I mean, we're talking about four times where it was a year ago
Nasdaq since this was said: ▲ +0.7% · 25,889 → 26,082 · as of 2026-09-10
Bullish
Visser says AGI is getting closer because coding keeps improving.
We are getting closer and closer to AGI because coding is getting better and better
Nasdaq since this was said: ▲ +0.7% · 25,889 → 26,082 · as of 2026-09-10
Bullish
Pompliano tells Visser AI demand is going exponential while compute supply is not keeping up.
This is a very weird market where demand is actually going exponential. And the supply is not responding in the way that it needs to
Nasdaq since this was said: ▲ +0.7% · 25,889 → 26,082 · as of 2026-09-10
Bullish
Visser says humanoid robots will be common in streets and many places within five years.
well, five years from now, we're going to have humanoids in the street, we're going to have humanoids in many, many places
Nasdaq since this was said: ▲ +0.7% · 25,889 → 26,082 · as of 2026-09-10
Bullish
Visser says AI is a bull market, not a bubble; the current mid-cycle slowdown means sideways action.
this is a bull market in AI. And there are going to be periods like now where I think we're in a mid cycle slowdown. That doesn't mean that it's a bubble. That just means they go sideways
Nasdaq since this was said: ▲ +0.7% · 25,889 → 26,082 · as of 2026-09-10
Bearish
Visser says hyperscalers have more trouble in the mid-cycle slowdown than AI infrastructure names like Micron and Marvell.
Some of the names have trouble, but I think the hyperscalers have more trouble than these guys do
Nasdaq since this was said: ▲ +0.7% · 25,889 → 26,082 · as of 2026-09-10
Bullish
Opening market take: 'I don't think there's anything fundamentally broken in the market' @1:17, though rates are higher, Middle East conflict continues and oil is high. She expects 'periods of consolidation before you can continue to ramp from here' @2:19 because physical buildout lags the agentic world.
“So you do have headwinds, but I I still think the fundamental story of a I and The margin expansion that we can eventually get from it productivity gains that we can get from it are true”
Nasdaq since this was said: ▲ +3.6% · 25,170 → 26,082 · as of 2026-09-10
Bearish
Todd 'Bubba' Horwitz tells Kitco today's tech/AI valuations look like the dot-com bubble, with warning signs everywhere, so serious pullbacks are due.
You know, this is like the dot-com bubble. We have some extreme valuations. We have warning signs galore out there for companies, for people, for the economy
Nasdaq since this was said: ▲ +1.6% · 25,679 → 26,082 · as of 2026-09-10
Bearish
Michael Oliver tells Kitco Nvidia already looks weak and is breaking key momentum numbers, so it could be a negative force on the market.
Nvidia's already looking weak to us. We're breaking certain key numbers right now on it that indicate it could provide a negative force.
Nvidia since this was said: ▲ +6.5% · 205.1 → 218.4 · as of 2026-09-10
Bullish
Thomas Laffont tells the All-In Summit that semiconductors are on a generational run that has massively outperformed the index since 2024.
Semis are on a generational run. I was just talking about this with my friend Brad Gerstner earlier.
Nasdaq since this was said: ▼ -2.8% · 26,831 → 26,082 · as of 2026-09-10
Bullish
Thomas Laffont tells the All-In Summit that AI's per-user memory needs could multiply several-fold, explaining the rally in memory-chip stocks.
the amount of memory per user could can tuple just based on the demand that these AI systems are requiring to provide their services. That helps explain why we've seen some of these moves in these memory companies.
Nasdaq since this was said: ▼ -2.8% · 26,831 → 26,082 · as of 2026-09-10
Bullish
Thomas Laffont tells the All-In Summit that AI ecosystem revenue is ~$140B today, will be ~$300B in 2026 and double in 2027.
We believe that it's about $140 billion today. It'll be about $300 billion this year, and it'll double in 2027.
Nasdaq since this was said: ▼ -2.8% · 26,831 → 26,082 · as of 2026-09-10
Bullish
Thomas Laffont tells the All-In Summit that AI-enabled ads at Meta and Google will go from ~25% to 100% of ads served, a $150B revenue pillar (no deadline).
We estimate currently that about a quarter of ads served by Meta and Google are AI enabled. We think that penetration will eventually go to 100%. That's $150 billion.
Nasdaq since this was said: ▼ -2.8% · 26,831 → 26,082 · as of 2026-09-10
Bearish
Maybe an AI bubble, a short-term duration mismatch
Andrew Ross Sorkin · CNBC / NYT · CNBC / '1929'Bullish
Gavin Baker (Atreides) tells All-In that LLM industry ARR across OpenAI, Anthropic, Gemini, Cursor, xAI and open source could reach $200-400B by end of 2026.
it's not hard to see 200, 300, $400 billion of ARR at the end of this year across all of the language models.
Nasdaq since this was said: ▼ -1.0% · 26,344 → 26,082 · as of 2026-09-10
Bullish
Gavin Baker tells All-In that GPU/AI-infrastructure spend will show strong ROI in 2026 on LLM token revenue alone, before counting ads and recommender use cases.
And it seems like there's going to be a really strong ROI this year, even excluding what are still some of the most economically important and profitable use cases for GPUs and AI infrastructure.
Nasdaq since this was said: ▼ -1.0% · 26,344 → 26,082 · as of 2026-09-10
Neutral
Gavin Baker tells All-In it is starting to look as if a CCP-funded campaign may be behind US anti-AI and anti-data-center sentiment (hedged speculation).
it is starting to feel or seem like there may be a CCP funded campaign against AI and data centers in America.
Nasdaq since this was said: ▼ -1.0% · 26,344 → 26,082 · as of 2026-09-10
Neutral
Gavin Baker tells All-In that AI valuations are cross-sectionally inconsistent: either Nvidia and memory re-rate up a lot, or richly valued power/cooling/optical names underperform.
If the multiples on the power, cooling optical names are correct, Nvidia, memory, they're going up a lot. If the multiples on Nvidia and memory are correct, everything else is probably going to underperform.
Nasdaq since this was said: ▼ -1.0% · 26,344 → 26,082 · as of 2026-09-10
Bullish
Gavin Baker tells All-In that with disaggregated accelerators older GPUs keep a 10-15 year useful life, so the two-year-GPU-life bear case on CoreWeave and hyperscalers is wrong and GPUs are financeable.
use Grok or Cerebrus for decode, and then those older GPUs, they have a useful life for 10 or 15 years, and this means that you can finance GPUs.
Nasdaq since this was said: ▼ -1.0% · 26,344 → 26,082 · as of 2026-09-10
Neutral
Gavin Baker tells All-In that rising rates and inflation are never good but AI fundamentals are getting much stronger, and flags that AI usage may be seasonal into summer.
rates going up and inflation going up is never good. But we have to hold what's happening with AI where the fundamentals are getting a lot stronger
Nasdaq since this was said: ▼ -1.0% · 26,344 → 26,082 · as of 2026-09-10
Neutral
Marc Benioff tells All-In that top enterprise software companies including HubSpot are trading at ~2x sales despite great quarters, a level he has never seen, and calls it a cyclical re-rating.
and the market's rerated. So you mentioned HubSpot, they're trading at two times sales. I've never seen that before. They just had a great quarter.
Nasdaq since this was said: ▼ -0.5% · 26,225 → 26,082 · as of 2026-09-10
Bullish
Marc Benioff tells All-In that the AI-kills-SaaS thesis is a market hypnosis not yet visible in the numbers; lots of enterprise software is still being sold.
a hypnosis around AI and, you know, we haven't seen it show up in the numbers yet. If it shows up in the numbers, maybe people will be right
Nasdaq since this was said: ▼ -0.5% · 26,225 → 26,082 · as of 2026-09-10
Bearish
Chamath Palihapitiya tells All-In that low-end SaaS is finished while large incumbents like Salesforce are safe because enterprise AI deployment is proving hard.
I think the low end of the market is basically finished. I think there is no safe space. I think the high end of the market, where Mark operates, where the large monoliths operate is quite safe.
Nasdaq since this was said: ▼ -0.5% · 26,225 → 26,082 · as of 2026-09-10
Bullish
Chamath Palihapitiya tells All-In that large-cap enterprise software is oversold and can re-rate upward; the opposite trade is incumbents with strong net dollar retention.
I think we're a little, I think we're a little oversold. And I now I think this consolidation and the rerating can happen in the opposite direction.
Nasdaq since this was said: ▼ -0.5% · 26,225 → 26,082 · as of 2026-09-10
Bearish
Chamath Palihapitiya tells All-In that once markets ask for the ROI on $3T of AI spend, AI-lab revenue multiples will come way down while incumbent software multiples recover.
So I think you're going to see these revenues, which are like way out of whack, like the multiples on revenue, multiples on assets, multiples of they're not even multiples of debt. Those will come way back down. And I think these go back up and you'll find a balance.
Nasdaq since this was said: ▼ -0.5% · 26,225 → 26,082 · as of 2026-09-10
Neutral
David Friedberg tells All-In that horizontal and vertical software stocks trade together today but should diverge, so there is probably an arbitrage (long horizontal / short vertical).
I also think that there's this kind of vertical, horizontal shift where they're kind of trading the same right now, but you could probably break them. You could break that trade. So I think there's probably a good arbitrage there.
Nasdaq since this was said: ▼ -0.5% · 26,225 → 26,082 · as of 2026-09-10
Bearish
Marc Benioff tells All-In that the 1000x-token-demand thesis is brainwashing: most tokens are wasted and a routing layer will send work to smaller models, so AI costs fall.
I don't think that's true. I don't think that that, I think we're getting brainwashed on that.
Nasdaq since this was said: ▼ -0.5% · 26,225 → 26,082 · as of 2026-09-10
Bullish
Huang says the $2T telecom base-station industry will in time become AI edge infrastructure.
So now, it's a $2 trillion industry. All of that in time will be transformed into an extension of the AI infrastructure.
Nasdaq since this was said: ▲ +18.1% · 22,091 → 26,082 · as of 2026-09-10
Bullish
Huang says physical AI lets the tech industry address a $50T industry for the first time.
Physical AI as a large category. It's technology industry's first opportunity to address a $50 trillion industry that has largely been void of technology until now.
Nasdaq since this was said: ▲ +18.1% · 22,091 → 26,082 · as of 2026-09-10
Bullish
Huang says digital biology/healthcare AI will inflect within five years.
In five years' time, I completely believe that the health care industry or digital biology is going to inflect.
Nasdaq since this was said: ▲ +18.1% · 22,091 → 26,082 · as of 2026-09-10
Bearish
Huang tells the hosts his main worry is that fear and doomerism slow AI adoption in the United States while other countries adopt it.
And so I'm just mostly worried about the diffusion of AI here in the United States.
Nasdaq since this was said: ▲ +18.1% · 22,091 → 26,082 · as of 2026-09-10
Neutral
Huang says compute demand rose about 10,000x in two years (generative to reasoning to agentic).
Now we're looking at in just two years, computation went up by a fact 10,000X.
Nasdaq since this was said: ▲ +18.1% · 22,091 → 26,082 · as of 2026-09-10
Neutral
Huang says helium could be a supply problem for chipmaking but the supply chain probably has buffer.
I think helium could be a problem, but it's also the case that the supply chain probably has a lot of buffer in it.
Nasdaq since this was said: ▲ +18.1% · 22,091 → 26,082 · as of 2026-09-10
Bullish
Huang says everything that moves will eventually be fully or partly autonomous.
We believe that everything that moves will be autonomous completely or partly someday
Nasdaq since this was said: ▲ +18.1% · 22,091 → 26,082 · as of 2026-09-10
Neutral
Huang says space data centers are exploratory and will take years; ground comes first.
And in the meantime, we're gonna explore what is the architecture of data centers look like in space? And it'll take years, it's okay.
Nasdaq since this was said: ▲ +18.1% · 22,091 → 26,082 · as of 2026-09-10
Bullish
Huang says every hospital instrument will become agentic.
Every single instrument, whether it's ultrasound or CT or whatever instrument we interact with in a hospital in the future, will be agentic.
Nasdaq since this was said: ▲ +18.1% · 22,091 → 26,082 · as of 2026-09-10
Bullish
Huang says robots will be everywhere within three to five years.
Three years to five years, we're going to have robots all over the place.
Nasdaq since this was said: ▲ +18.1% · 22,091 → 26,082 · as of 2026-09-10
Bullish
Huang says China's micro-electronics, motors, rare-earth and magnet ecosystem is the world's best and the global robotics industry will depend on it.
I think China is formidable, and the reason for that is because they're micro-electronics, they're motors, they're rare earth, they're magnets, which is foundational to robotics.
Nasdaq since this was said: ▲ +18.1% · 22,091 → 26,082 · as of 2026-09-10
Bullish
Raoul Pal tells Mando the AI race is the most important thing humanity has done and the most powerful (probably last) technology it will invent.
this AI race is the biggest race in all of humanity
Nasdaq since this was said: ▲ +10.2% · 23,671 → 26,082 · as of 2026-09-10
Bullish
Gerstner cites OpenAI's reported $13B 2025 revenue against ~$1.4T of multi-year compute commitments ($500B NVIDIA, $300B AMD/Oracle, $250B Azure as read on tape). Altman: 'We do plan for revenue to grow steeply.' @13:05 and 'I would love to tell them they could just short the stock and I would love to see them get burned on that.' @13:40.
“We're doing well more revenue than that. Second of all, Brad, if you want to sell your shares, I'll find you a buyer. Enough.”
Nasdaq since this was said: ▲ +10.6% · 23,581 → 26,082 · as of 2026-09-10
Bearish
Continues '...probably several points along the way.' Asked whether he agrees with Jensen Huang that a compute glut is a near-zero chance in 2-3 years, Altman instead warns of cycle risk: 'if a very cheap form of energy comes online soon at mass scale and a lot of people are going to be extremely burned with existing contracts they've signed.' @19:53; 'Some people are going to get really burned.' @20:33. He notes cost per unit of intelligence has been falling ~40x/year, 'a very scary exponent from an infrastructure build out standpoint.'
“There will come a glut for sure. And whether that's like in two to three years or five to six, Satya and I can't tell you, but like it's going to happen at some point”
Nasdaq since this was said: ▲ +10.6% · 23,581 → 26,082 · as of 2026-09-10
Bullish
Gerstner relays Jensen Huang's earlier BG2 answer as the premise for asking Altman and Nadella if they agree; Nadella redirects to power constraints ('the biggest issue we are now having is not a compute glut, but it's a power' @18:26) and Altman gives his glut-for-sure answer. Indirect speech, not a verbatim Huang quote on this tape.
“I asked Jensen on this pod, if there was any chance over the course of the next five years, we would have a compute glut. And he said it's virtually non-existent chance in the next two to three years.”
Nasdaq since this was said: ▲ +10.6% · 23,581 → 26,082 · as of 2026-09-10
Bullish
Asked about compute-glut risk: 'the biggest issue we are now having is not a compute glut, but it's a power' @18:26. Later, on shaping Azure's book: 'we're not demand constrained, we're supply constrained.' @45:53. Confirms Azure (39% growth, $93B run-rate as stated on tape) could have grown 41-42% with more compute: 'Absolutely. There's no question.'
“And in fact, that is my problem today, right? It's not a supply issue of chips. It's actually the fact that I don't have warm shelves to plug into.”
Nasdaq since this was said: ▲ +10.6% · 23,581 → 26,082 · as of 2026-09-10
Bearish
Nvidia makes 100%+ of AI profits; feels close to the 1999 bubble
Peter Thiel · Founders Fund · All-In PodcastOn the scoreboard
Voices in this debate who also have scored calls on the ledger.
Head to head
Pick any two voices in this debate; their freshest claims go side by side.
Follow The AI Trade
One short email when this debate moves: new calls, reversals, verdicts. Plain words, free.
Before you trust the call, check the record.
We write down what public figures predict, in their exact words, and check the result when the deadline arrives. Every line links to its source. How we score · Corrections · The scoreboard · The data
© 2026 Pundit Rumble · independent and unaffiliated · updated every market morning · not investment advice.
We write down what public figures predict, in their exact words, and check the result when the deadline arrives. Every line links to its source. How we score · Corrections · The scoreboard · The data
© 2026 Pundit Rumble · independent and unaffiliated · updated every market morning · not investment advice.