Nvidia Fallout, AI Financing Risk and Apple’s Safe-Haven Surge 11/12/26

NVIDIA

Across the shows: 131 speakers in other episodes, 117 bullish and 84 bearish calls, past 30 days

• Guy Adami · CNBCUncertain

  Nvidia may have found a tradable bottom at prior support, but more downside shoes will drop.

  Adami notes the stock hit a low near the prior support, which marks a possible tradable bottom. He cites the January 2025 high near 148 that cascaded down to about 91 in spring 2025, with today's low at past resistance now support. He still sees many more shoes to drop, and says the structure of the AI trade built on Nvidia is starting to crack.

  “So I think to Dan's point, there might be a tradable bottom here for sure. But I think the overriding point, in my opinion, there are many more shoes to drop in this entirety of this thing.”at 8:00

• Tim Seymour · Seymour Asset ManagementBullish

  Nvidia won't be replaced in the short run; customers still want every chip it makes.

  Seymour argues that in the short term Nvidia chips will keep going where they are going. The bigger question is the relationships, which are priced in. He says there is a sense that everybody wants every last chip Nvidia has and will keep buying. So he does not think Nvidia can be replaced soon, even if long-term risks remain.

  “right now in the short term, I think NVIDIA chips are going where they're going to go... I don't think you're replacing Nvidia in the short run.”at 9:08

• Brian Kelly · BKCM LLCBearish

  Anthropic's AMD chip work threatens Nvidia's lead as AI redesigns how it operates.

  Kelly points to Anthropic's announcement that Claude taught itself to use AMD chips over Nvidia's. He argues this will create an acceleration effect, with more intelligence per chip as AI redesigns how it operates. That threatens Nvidia's lead, even though the stock is already down a lot.

  “Anthropic announced that Claude had taught itself how to use AMD chips over NVIDIA. So, I mean, when you're talking about this acceleration effect... I think NVIDIA's lead was certainly threatened.”at 10:40

• Dan Nathan · RiskReversal MediaBearish

  Nvidia's sales-diversion and allocation issues won't go away; he's staying out after the rally.

  Nvidia still earns about 73% gross margins versus 55% at AMD, on about $400 billion of sales against AMD's $52 billion. Nathan says customers were taken on price, and Nvidia has been aggressive about allocations. Nvidia likely sent chips to Singapore that were going to China, while US national champions could not get the allocations they wanted. He has seen the stock rally 50% in a week, and he thinks the diversion issue will persist, so he does not want to trade it.

  “NVIDIA has been savage about allocations... when you think about all the ways in which they look to have covered up the ways in which their chips were getting over there, To me, this is not something that's going away.”at 12:09

• Gil Luria · D.A. DavidsonBullish

  Nvidia should still rise next year despite probe risk; he kept his Buy rating.

  Luria says the probability weight still favors Nvidia delivering next year's numbers, at which point the stock will be higher. Even if there was wrongdoing or Asia sales are limited, Nvidia can probably still sell the chips it planned for next year and reach the same numbers. The main risk is more limited Asia sales. That is why he kept the Buy but deprioritized the stock in favor of AMD, Broadcom and Micron.

  “The probability weight of outcomes is still that they are able to deliver the numbers for next year, at which point the stock will be higher.”at 15:02

• Gil Luria · D.A. DavidsonUncertain

  Nvidia cannot miss its upcoming quarter; a miss means another big leg down.

  Luria argues the upcoming quarter is the most important in a long time because the company has to show up. It has to make the numbers and reiterate guidance into next year, or the stock falls further. He says there is no way Nvidia can miss without a big drop.

  “There is no way they can miss. If they miss, there's another big leg down.”at 17:09

Cerebras Systems

Across the shows: 3 speakers in other episodes, 1 bullish and 1 bearish calls, past 30 days

• Brian Kelly · BKCM LLCBullish

  Inference demand will grow far faster than other data center uses, making Cerebras an attractive play.

  Kelly argues inference will grow a lot more than the rest of data center demand, and the hyperscalers are already buying it. He prefers going down the stack to chips like Cerebras and Cisco rather than training-chip names. He acknowledges that valuation caught up with the stock during the tech selloff.

  “I mean, inference is going to grow a lot more. We talked about the hyperscalers. I think this is one way to do it.”at 27:27

• Dan Nathan · RiskReversal MediaBullish

  Cerebras could be a rocket ship; cash-rich, near profitability, and growing sales near 100%.

  Nathan says roughly a third of Cerebras's market cap is in cash, and it is expected to be profitable next year. Its sales are growing close to 100% year over year. He sees an opportunity to diversify away from the training-chip names, and a Microsoft customer win makes it more attractive.

  “Stock could be a rocket ship right here... I think maybe a third of their market cap is in cash. They're expected to be profitable next year. They're growing sales, I want to say, 100% year over year, too.”at 27:47

Bitcoin

Across the shows: 25 speakers in other episodes, 22 bullish and 8 bearish calls, past 30 days

• Brian Kelly · BKCM LLCBullish

  Bitcoin may bounce from around $67,000 after a selloff that Master news accelerated.

  Bitcoin has sold off sharply, and Master news accelerated the decline toward roughly $67,000. Kelly sees that level as a possible bottom, and he will watch for a bounce there. His view is tactical and near term. He is unsure whether the growth of on-chain finance will benefit the Bitcoin token itself or only related equities.

  “it somewhere around maybe the $67,000 level to me looks like, Hey, maybe this thing bounced. That's where I'm going to be looking to see if there's a bounce.”at 42:01

• Tim Seymour · Seymour Asset ManagementBullish

  Bitcoin is distinct from worthless tokens and can probably go higher.

  Seymour thinks many crypto assets are worthless, but he separates Bitcoin from that group. He thinks Bitcoin probably can go higher. His reasoning is brief. He points to the way large banks, including JPMorgan, are moving onto the blockchain rails that Kelly describes as the transition.

  “I think a lot of crypto assets are garbage. I'm not going to say that about Bitcoin. I think it probably can go higher, but there's a lot of tokens out there that are worthless.”at 43:29

Okta

Across the shows: 1 speaker in other episodes, 1 bullish and 0 bearish calls, past 30 days

• Brian Kelly · BKCM LLCBullish

  Okta benefits from exponential growth in onboarding AI agents at companies.

  Kelly describes Okta as a company that onboards new employees and gives them credentials. He says it now has exponential growth onboarding AI agents at many companies. He sees the agentic economy as an exponential growth curve that is just beginning, and says Okta is a better play than worrying about chip share.

  “...like Okta, O K T A, right? This is a company that basically onboards new employees... But now they've got exponential growth onboarding agents at all these companies.”at 13:06

Advanced Micro Devices

Across the shows: 20 speakers in other episodes, 21 bullish and 1 bearish calls, past 30 days

• Gil Luria · D.A. DavidsonBullish

  AMD is the priciest of the AI chip trio but has upside from execution.

  Luria concedes AMD is the most expensive of AMD, Broadcom and Micron. He attributes the premium to strong execution and upside to its numbers. He still lists AMD among the better ways to play AI demand, so he sees the premium as earned.

  “...AMD is the priciest of the three... It's because they've executed so well and there's upside to numbers.”at 16:07

Broadcom

Across the shows: 14 speakers in other episodes, 14 bullish and 6 bearish calls, past 30 days

• Gil Luria · D.A. DavidsonBullish

  Broadcom is a better way to play AI data center build-out, trading at a below-market multiple.

  Luria says AMD, Broadcom and Micron are a better way to play the same trend of growing AI demand that translates into data center build-out. Broadcom still trades at a much lower multiple than AMD, and below the market multiple. Together with Micron and AMD it can replicate Nvidia's profit pool at a blended multiple below the market.

  “...they're still trading at a much lower multiple... and they will participate in the data center build out.”at 16:07

Micron Technology

Across the shows: 24 speakers in other episodes, 23 bullish and 16 bearish calls, past 30 days

• Gil Luria · D.A. DavidsonBullish

  Micron trades at a single-digit multiple while still benefiting from AI data center build-out.

  Luria argues Micron is a cheaper way to play growing AI demand that drives data center build-out. He notes Broadcom and Micron trade at below-market multiples, with Micron at a single-digit multiple. Micron will participate in the build-out while carrying far less valuation risk than AMD.

  “But Broadcom and Micron still trading at either a below market multiple or a single digit multiple in the case of Micron.”at 16:21

Eli Lilly and Company

Across the shows: 13 speakers in other episodes, 12 bullish and 2 bearish calls, past 30 days

• Tim Seymour · Seymour Asset ManagementBullish

  Investors don't care much about Lilly's valuation, favoring top-quality names.

  Seymour says the market is rotating into the highest-quality names, and Lilly is one of them. Lilly hit an all-time high on approval of the first GLP-1 weight-loss pill in China, and it is up 33% this year. He thinks investors don't care much about Lilly's valuation.

  “I don't think people care that much about valuation at Lilly.”at 32:35

OpenAI

Across the shows: 42 speakers in other episodes, 22 bullish and 33 bearish calls, past 30 days

• Gary Marcus · New York University (professor emeritus)Bearish

  OpenAI may be the WeWork of AI; a failed IPO would unwind circular financing.

  WeWork talked about going public, made several false starts, and then collapsed without listing. Marcus has long thought OpenAI could follow that path. If the planned IPOs of OpenAI and Anthropic fall apart, the circular financing tying the AI ecosystem together could unravel and spread contagion across the sector.

  “If either of the IPOs fall apart, it's all over. And I think you've heard me say before that I've always thought that OpenAI might be the WeWork of AI.”at 38:03

Anthropic

Across the shows: 116 speakers in other episodes, 164 bullish and 91 bearish calls, past 30 days

• Gary Marcus · New York University (professor emeritus)Bearish

  Anthropic could also prove a WeWork-style IPO failure, unwinding circular AI financing.

  Marcus extends his WeWork analogy from OpenAI to Anthropic and says it could turn out the same way. He sees the planned IPOs as critical to the circular financing among AI companies. If either IPO collapses, he expects a contagion effect. The argument is hedged and speculative.

  “Could turn out that, [Anthropic] is too.”at 38:05

Nscale

Across the shows: 2 speakers in other episodes, 0 bullish and 2 bearish calls, past 30 days

• Tim Seymour · Seymour Asset ManagementBearish

  nScale's capital intensity and lack of profitability make it a risky AI infrastructure name.

  Seymour sees the AI buildout as highly capital-intensive. nScale is his example of a company that is unprofitable and needs large amounts of funding. He adds that it is only about 5% of the way to the capacity it has contracted to deliver, so the funding gap is large relative to its present position.

  “I understand that there is a need for capital. The capital intensity going on here and again, N scale is a perfect example. Lack of profitability.”at 38:40

Figure AI

• Brian Kelly · BKCM LLCBullish

  Figure's on-chain HELOC lending and 11% earnings growth make it a bullish fintech name.

  Figure takes home equity lines of credit and puts them on chain, which Kelly sees as an early example of finance moving onto blockchain rails. The company reported strong earnings that day, with revenue or earnings up 11%. Kelly expects the company to extrapolate that growth as more assets move on chain.

  “And you look at a name like today, like Figr... They basically are taking HELOCs and putting them on chain. They had tremendous earnings today, up 11%, and they're going to extrapolate that.”at 42:04

Apple

Across the shows: 51 speakers in other episodes, 64 bullish and 42 bearish calls, past 30 days

• Tim Seymour · Seymour Asset ManagementBullish

  Apple is no more expensive than a year ago and looks more interesting at this multiple.

  Seymour rejects the Bernstein argument that treats Apple like Costco or Walmart, calling that comparison unfair. He says the stock is not more expensive than it was a year ago and that the current multiple makes it more interesting. He also argues investors underrate Apple's ability to monetize through its App Store and margins.

  “I don't think the company is any more expensive than it was a year ago. In fact, I think it's more interesting at this multiple than it was a year ago.”at 45:31

AI infrastructure & data centers

• Tim Seymour · Seymour Asset Management

  Higher yields make neocloud funding harder and costlier, tapping the brakes on AI buildout.

  Seymour argues that yields are rising, which makes funding for all of this harder to obtain. Neocloud capital is still available, but it comes at a higher price. He thinks this taps the brakes on the AI trade, and that rising rates are a big story for AI.

  “I mean, yields are going higher, so funding for all of this stuff, funding for neo clouds is getting a lot harder in terms of where the capital is coming from, but it's coming at a higher price. I think all of this kind of taps the brakes.”at 4:27

• Dan Nathan · RiskReversal Media

  Data center build-out is slowing on power and cooling limits; investors should favor AI users over infrastructure.

  Nathan says data centers were already challenged by power and cooling constraints that were going to slow the build-out. He concludes investors should stay away from the AI infrastructure trade. He prefers the companies that use AI, because the infrastructure build-out is decelerating.

  “We know that data centers were challenged because of power and cooling and a lot of different things that was going to slow that build out. So for me, I think you just stay away from the AI infrastructure trade and you go back to the AI, the people that are using AI.”at 5:44

• Gary Marcus · New York University (professor emeritus)

  Neocloud AI companies are on a fragile financing edge; Nvidia trouble will raise credit risk.

  AI infrastructure companies need continuous outside money to keep building. Marcus cites CoreWeave and nScale as examples where problems already appeared. He expects Nvidia's trouble to push credit spreads higher because investors will doubt whether the buildout economics work. If one of these companies fails, he argues the circular financing could spread contagion to valuations across large language model companies and to lending banks, possibly prompting too-big-to-fail bailouts.

  “They need things to keep it going. These companies are all just on a fragile edge.”at 36:26

• Dan Nathan · RiskReversal Media

  Neocloud funding gaps could cancel non-cancelable orders and snowball into frontier-lab risk.

  EnScale says it has non-cancelable orders but needs capital to build the compute to fulfil them. If it cannot build out, those orders are cancelled. Nathan asks whether this small case could snowball into much bigger ones. He notes that Nvidia is a single point of failure on the hardware side, and a frontier lab could play the same role on the demand side.

  “These are non cancelable orders that they have. They need the capital to fulfill those orders. If they can't build out the compute, then the orders that they have get canceled.”at 37:34

• Brian Kelly · BKCM LLC

  Data center build costs are rising from community demands and banks pulling back on loans.

  Kelly expects the cost of building data centers to rise. Politicians will require more community benefits, possibly including more solar. Lenders, which he refers to as banks, are likely pulling back on many of their loans. He also mentions the possibility of a too-big-to-fail bailout for AI companies, but says he is not claiming that is happening. Higher costs are a problem for AI growth, which he says depends on this buildout.

  “So the cost to build these things are going to go up, which is going to be a problem.”at 40:38

US Treasuries & yields

• Tim Seymour · Seymour Asset Management

  He is short two-year notes, betting rates keep rising from the October uptrend.

  Seymour cites the uptrend in rates going back to October of last year. Despite volatility over the last month, that trend has held. He reads recent Waller comments and the possibility of a 50 basis point hike as the Fed trying to restore credibility. He says he is short two-year notes, betting rates will rise.

  “I think interest rates... that uptrend going all the way back to October of last year... It's held, and I think they do hold. Yeah, I mean, I'm short of two year notes, right? So I'm betting that interest rates are going to be going higher here.”at 22:25

• Brian Kelly · BKCM LLC

  Two-year and ten-year yields likely rise further on global growth and inflation pressure.

  Kelly says the two-year rate looks like it wants to go much higher, and the ten-year rate will go higher too. He points to global growth, including stronger UK GDP attributed to AI. He adds the inflation factor, citing oil and the dollar. Together these support higher rates.

  “Anyway, back to the two year rates. They actually look to me like they do want to go much, much higher here. Even the ten year rate goes higher.”at 23:07

AI: overall outlook

• Tim Seymour · Seymour Asset Management

  The AI trade still works despite Nvidia's turmoil.

  Seymour says the market since Monday morning has shown the AI trade still works. He points to the pattern across the semiconductor complex, where parts of the trade are working even as Nvidia sold off. He treats the Nvidia issue as company-specific rather than a broad AI signal.

  “I think what we've seen from the market since Monday morning is that the AI trade absolutely still works.”at 4:27

US dollar

• Tim Seymour · Seymour Asset Management

  The dollar at 18-month highs may break higher, hurting investors in global stocks.

  Seymour notes the dollar is at 18-month highs and looks like it wants to break higher. That is a headwind for investors who buy global stocks. He ties this to the rate picture, since higher US yields support the dollar.

  “Look at that dollar that's at eighteen month highs and looks like it wants to break higher and that's not great for the guy that likes to invest in global stocks.”at 22:11

US growth & recession risk

• Dan Nathan · RiskReversal Media

  The US economy looks to be teetering, with home sales at 16-year lows and slow growth.

  Nathan cites existing home sales down 5.5% from the prior month, at 16-year lows, last seen around the financial crisis. He says the consumer has had problems and the US is growing more slowly than inflation. He says the US economy seems to be teetering, and AI productivity is needed to support it.

  “...You have this existing home sales number down 5.5% from the prior month, down at sixteen year lows... So it seems like The U. S. Consumer or U. S. Economy really does seem to be teetering a little bit.”at 23:31

Pharma & biotech

• Guy Adami · CNBC

  Healthcare still looks attractive, with valuations not stretched like other sectors.

  Adami says healthcare has been bid up but not to valuations that don't make sense, unlike some other sectors. Healthcare names like J&J and Merck are near all-time highs. If the AI trade slows because of Jensen Huang's departure, healthcare could keep attracting money as a rotation trade, as it did in October.

  “I still like healthcare here. I mean, if you are to believe that the AI trade will in fact slow because of Jensen Huang's departure, then health care could maybe continue.”at 31:52

AI bubble & tech valuations

• Gary Marcus · New York University (professor emeritus)

  Circular AI financing could collapse quickly, since the LLM companies lose money and depend on it.

  Marcus says Jensen Huang was a big risk taker who put money into many companies. His replacement is a more conservative CFO, who may question these bets as confidence in AI falls. All the large language model providers are losing money and need lots of investment, so the circular financing could collapse very quickly. He cites Anthropic's IPO, where about $10 billion is supposed to come from Nvidia.

  “And so the circular financing thing could collapse, and it could collapse very quickly... all of the LLM companies, the large language model providers, are losing money, and they need lots of investment to sustain themselves.”at 33:49

Stablecoins & crypto policy

• Brian Kelly · BKCM LLC

  Finance moving on chain will replumb the financial system into a huge growth curve.

  Kelly sees the big trend as finance moving on chain. Lending, securities and other assets will be tokenized, which will rebuild the financial system and let people do things that were not possible before. He calls this a huge growth curve. He is unsure whether the value from this growth flows to tokens like Bitcoin, Ethereum or Solana or to listed equities, which is why he prefers securitized and equity exposure.

  “to me, what the big trend here is all of finances coming on chain. You're going to replumb the entire financial system.”at 42:04