The David Lin Report · · 53 min
Nvidia Crushes Earnings: Will Tech Stocks Rebound Or Crash More? | Jay Singh
With Jay Singh
ChorusBring the market's conversation to your AI. Learn more
Summary
Jay Singh joins David Lin to dissect Nvidia's latest quarter, which beat on revenue and EPS and raised next-year revenue growth outlook from 70% to 75% even with China excluded from guidance. He argues Nvidia's roughly $400B of off-balance-sheet guarantees to neoclouds and data-center projects are manageable, with about 18% loan-to-value on his math, though he flags the risk that some backstops could be exposed. He is bearish on the leadership overhang from Jensen Huang's departure and sees the market pricing a founder premium, while calling the 10x forward multiple cheap relative to 2028 earnings. On AI labs, he sees Anthropic's large reported loss as mostly non-cash and its revenue growth as strong, but warns that cheap open-source models from China will capture a large share of usage and pressure margins at both Anthropic and OpenAI, with OpenAI also facing cash burn and SoftBank funding risk. The discussion then covers a 'pop and drop' pattern in recent mega IPOs, with Cerebras as the example and an Anthropic offering as a warning, shorting CoreWeave as a commodity neocloud, a crowded defensive rotation and a stressed tape, and a Fed that he expects to hike in December.
Key takeaways
- Jay Singh (uncertain on NVIDIA): Nvidia's roughly $118B of guarantees to data-center builds could be at risk as its CFO pulls back on neocloud support.
- Jay Singh (bullish on NVIDIA): Nvidia's leverage is very low, so it could honour all its guarantees even if the stock fell 50%.
- Jay Singh (bullish on NVIDIA): Nvidia's raised next-year growth outlook to 75% shows demand beyond the restricted countries.
- Jay Singh (bearish on Anthropic): Cheap open-source models from China will take most US queries from Anthropic and OpenAI, a major risk.
- Jay Singh (bearish on CoreWeave): CoreWeave doesn't have a long-term reason to exist; neoclouds are commodity businesses, so he is short.
What was said, by company and theme
25 views from 1 speaker. Every quote is checked against the episode's transcript (how).
NVIDIA
Across the shows: 178 speakers in other episodes, 182 bullish and 182 bearish calls, past 30 days
Where everyone stands on NVIDIA
Jay Singh · Special Situations ResearchBearish
Nvidia's stock sold off because markets see Jensen Huang as irreplaceable, and his exit carries a large premium loss.
Huang is a founder and innovator, not a normal CEO. He has 65 direct reports, speaks directly with customers, does the deal-making, and is essential to the whole AI ecosystem. The chairman said on the call that Huang's return is not being considered. Singh estimates the stock has given up over 10%, about $500B, since the affair was disclosed and the Mega Speed story emerged. Shares fell to about $1.55 after hours before bouncing. Singh says the market is pricing a Jensen Huang premium, and he sees governance concerns, such as the board reportedly lacking D&O insurance, but does not weigh the spy allegations as proven.
“this is not a normal CEO, this is, you know, founder and innovator. He has 65 direct reports. He speaks directly with customers ... and is essential to the entire AI ecosystem.”at 7:41
Jay Singh · Special Situations ResearchBullish
Nvidia's stock at about 10x forward earnings looks cheap, with growth affirmed on the call.
Nvidia trades at about 10 times forward earnings. Growth had been questioned before the call, but management affirmed it. Singh says the multiple has compressed because the market is not trading on 2027 earnings. It is trading on 2028 earnings, and Nvidia is sold out through 2027. That framing implies the low multiple reflects uncertainty about the later period rather than weak current earnings.
“the stock is trading at 10 times forward earnings. Now the E, you know, that that, growth, was under question before this call, but it looks like, you know, they've affirmed the growth.”at 8:00
Jay Singh · Special Situations ResearchBearish
Gross margin guided down from 74.9% to about 71.5% is the only red flag in Nvidia's numbers.
Singh reviewed Nvidia's quarter and found the numbers quite good overall. The one red flag he saw was that gross margin is guided to fall from 74.9% to about 71.5% in the fourth quarter. He treats this as a modest concern against an otherwise strong report, which had a double beat and raised growth expectations.
“The only, you know, red flag I saw in the actual numbers was that gross margin would be falling from 74.9% to about 71.5% for the fourth quarter.”at 8:07
Jay Singh · Special Situations ResearchUncertain
Nvidia's roughly $118B of guarantees to data-center builds could be at risk as its CFO pulls back on neocloud support.
Nvidia is effectively the bank for neoclouds and frontier labs. It holds about $62B of equity investments (e.g., Intel, which it has been trimming), about $118B of guarantees including credit support for an SB Energy data-center campus in Ohio, and about $28B of guarantees to neoclouds. Its accounts receivable hit a record $81B and inventory reached $40B. Historically the CFO pushed back on neocloud guarantees, and she said on the call she would be disciplined and selective. Singh's worry is that this means some of the guarantees could be at risk. He notes Nvidia only backstops once a site is operating and under construction, not before permits or land are secured.
“The thing that I'm worried about on the NeoClouds is that Colette historically had pushed back when it came to these NeoCloud guarantees, and she said she would be disciplined and selective on supporting the NeoClouds, which I think means that some of the guarantees could be at risk.”at 8:46
Jay Singh · Special Situations ResearchBullish
Nvidia's leverage is very low, so it could honour all its guarantees even if the stock fell 50%.
Singh uses an LTV-style test. Nvidia has a roughly $3.8T market cap, $99B of cash and $38B of debt, so it starts with no net debt. Adding the $227B existing buyback authorization plus a new $150B buyback, assumed fully executed immediately, gives about $300B of net debt. Adding about $400B of off-balance-sheet guarantees gives about $700B in total, or an LTV of roughly 18%. Its debt has not risen much, and its net cash position this year is better than two years ago. Unlike other tech companies that are spending on GPUs, Nvidia is the company benefiting from that spending, so it has the firepower to honour its commitments.
“this company is a very very low LTV. So yes, NVIDIA does... Even if the stock went down another 50% from today, they actually would have the firepower to honor all their guarantees.”at 15:10
Jay Singh · Special Situations ResearchBullish
New US export curbs on six countries hurt Nvidia's sales only marginally, since other customers absorb the supply.
The Commerce Department now requires US permission before companies in China, Malaysia, Thailand, Singapore, Indonesia and Brazil can buy Nvidia's advanced AI chips. Singh says the rule partly reflects the Jensen Huang scandal. He argues the diversion was known for over a year, since Singapore physically received only about 1% of Nvidia sales and acted as an accounting hub for shipments to China and elsewhere. Nvidia says it is closing loopholes, and China contributed $2.5B this quarter, which was excluded from guidance. Management still raised its outlook, so the restrictions have a limited impact.
“Of course, it has an impact on chip sales abroad to those six countries, but as per the call, they're well aware of these restrictions and they spoke about these restrictions.”at 18:20
Jay Singh · Special Situations ResearchBullish
Nvidia's raised next-year growth outlook to 75% shows demand beyond the restricted countries.
Nvidia beat on both revenue and EPS. On the call it raised its revenue growth outlook for next year from 70% to 75%, even after saying China sales are excluded from guidance. Singh reads this as a sign there are enough customers worldwide to absorb supply that would have gone to the six restricted countries. Going into the call, Nvidia was sold out through 2027. The stock rose because management said next year would grow more than expected. The forecast is company guidance rather than a number Singh commits to.
“They actually increased the company's revenue growth outlook for next year from 70% to 75%. ... So what they effectively said on the call indirectly is that there's more than enough customers in the world to absorb the excess supply from these six countries.”at 18:30
Anthropic
Across the shows: 153 speakers in other episodes, 269 bullish and 133 bearish calls, past 30 days
Where everyone stands on Anthropic
Jay Singh · Special Situations ResearchBullish
Anthropic's $104B nine-month loss is mostly non-cash, and its roughly $100B liquidity means no near-term funding issue.
Most of the reported loss is accounting noise. About $72.4B of the first-half net loss of $80.7B was a fair-value remeasurement of convertible notes, driven by the company's rising valuation, and about $8.8B came from other non-cash financing and derivative adjustments. Of the nine-month loss of $104B, most was non-cash, so the underlying loss is closer to $10B. With roughly $100B of liquidity, Singh does not expect a liquidity problem this year. He also notes revenue growth is strong, rising from about $1.16B in the first half of 2025 to about $16B in the first half of 2026, with an 85% gross margin and about 80% of revenue from usage-based consumption.
“was non cash, so it's really a 10,000,000,000 loss. And given the, you know, 100,000,000,000 odd liquidity they have, I don't think they're gonna have a liquidity issue, this year.”at 34:51
Jay Singh · Special Situations ResearchBearish
Cheap open-source models from China will take most US queries from Anthropic and OpenAI, a major risk.
Singh estimates US consumers will use Anthropic and OpenAI for only about 30% of their queries and coding, with the rest going to open-source models that were started in China. He cites the cost gap: Microsoft employees use open-source models because tokens cost about 95% less. He argues Anthropic and OpenAI want to keep people off these models because they would hurt margins, which he sees as the biggest risk facing both companies. He calls it the central business risk, alongside safety, hacking and concentration risks.
“US consumers will only use Anthropic and OpenAI for like 30% of their queries and coding and the rest will go to these open source models that were effectively started in China. Even Microsoft employees use open source models because the tokens cost 95% less.”at 35:25
Jay Singh · Special Situations ResearchBearish
Anthropic faces safety, customer-concentration and governance risks that raise public-oversight concerns.
Singh lists several risks for Anthropic. Some of its models can resist shutdown and conceal information, and there are hacking risks and runaway capability thresholds. Its top five customers provide 25% of revenue. Seven founders retain 50.1% of the voting power, which he says raises public-oversight concerns. He cites these as a reason for caution rather than giving a detailed quantified view.
“some of these models can resist shutdown, conceal information, the risks around hacking, runaway capability thresholds, revenue concentration. ... Seven founders retained 50.1% of the voting power, which raises public oversight con... Concerns.”at 36:16
Jay Singh · Special Situations ResearchBearish
Anthropic will burn tens of billions and need more capital within two years; its IPO may sell off sharply.
Jay discusses the upcoming IPO with an offer price range of $70.78 and warns that the stock could rally above $100 and then sell off very sharply, given the uncertain backdrop for AI and the fact that 41% of data centers are being built in blue states. He says the company will burn cash, though not the roughly $100 billion loss that appears once non-cash items are excluded. It will still burn tens of billions of dollars and will likely need to raise more capital within the next two years. He follows this with the lockup unlock schedule.
“And this company will burn cash... It's not gonna be anywhere near the 100,000,000,000 loss because of the non cash, but they will be burning tens of billions of dollars, and they will need to likely raise more capital within the next two years.”at 41:18
OpenAI
Across the shows: 42 speakers in other episodes, 20 bullish and 33 bearish calls, past 30 days
Where everyone stands on OpenAI
Jay Singh · Special Situations ResearchBearish
Open-source models from China will take most US queries from OpenAI, a major business risk.
Singh expects US consumers to use Anthropic and OpenAI for only about 30% of their queries and coding, with the rest going to cheaper open-source models started in China. Microsoft employees already prefer them because tokens cost about 95% less. He says Anthropic and OpenAI do not want people using these models because it would hurt their margins, so this competition is the biggest business risk for OpenAI.
“US consumers will only use Anthropic and OpenAI for like 30% of their queries and coding and the rest will go to these open source models that were effectively started in China.”at 35:25
Jay Singh · Special Situations ResearchBearish
OpenAI's cash burn is a much bigger risk than Anthropic's, and SoftBank may not fund it.
Singh says OpenAI's cash flow burn is a much bigger risk than Anthropic's. On top of that, SoftBank, one of OpenAI's biggest investors, may not have enough money to keep funding it. He presents this as a financing risk separate from the open-source competition risk.
“OpenAI's cash flow burn is a much bigger risk than [Anthropic], and OpenAI also has the issue of SoftBank not having enough money to continue funding them given they're one of the biggest investors.”at 36:42
Cerebras Systems
Across the shows: 5 speakers in other episodes, 3 bullish and 1 bearish calls, past 30 days
Where everyone stands on Cerebras Systems
Jay Singh · Special Situations ResearchBearish
Cerebras's IPO has been a disaster; the stock keeps falling through lockup unlocks and earnings.
Cerebras (CBRS) IPO'd in May 2026 at $1.85 and raised about $6.38 billion. It rallied 68% to $3.11, then fell more than 40% below its IPO price and 67% off the first-day close within a few months. It fell another 21% around its November quarterly earnings and unlock. Jay offers this as a clear example of the pop-and-drop IPO pattern and of why the stock is a poor place to hold after listing.
“I would say that recent mega IPOs, most notably like Cerebras, which has been a disaster with several unlocks”at 38:26
StandardAero
Jay Singh · Special Situations ResearchBullish
StandardAero is a high free-cash-flow aftermarket aircraft parts business he found attractive; the stock rose about 50%.
Jay's firm bought StandardAero (SARO) near $20 in October and sold it at about $29, a gain of almost 50%. He liked it for its aftermarket aircraft parts business, which serves commercial and government aviation and power customers and includes missile defense systems. He says it is easy to forecast revenue and converts a high share of earnings into free cash flow. Jay compares it to FTAI. He says the firm is now fully out of the position.
“the reason why was this company manufactures aircraft parts and equipment and missile defense systems serving commercial and government aviation and power customers. So we thought this was very interesting name.”at 45:16
Entravision Communications
Jay Singh · Special Situations ResearchBullish
Entravision's fast-growing advertising business makes it a mid-cap media name the market overlooks.
Entravision (EVC) has moved from being a TV company to a diversified media company with a growing business. Jay's firm bought it around $6 a share because its advertising business was growing very quickly. The stock rallied to about $9 within a month after earnings, a roughly 50% gain. He argues mid-cap companies like this are under-followed. Jay describes the position as a purchase, and the transcript does not say whether it was sold.
“it's a diversified media company with a growth business and we were buying this around $6 a share because its advertising business was growing so rapidly.”at 47:15
Redwood Trust
Jay Singh · Special Situations ResearchBullish
Redwood's baby bonds at about 80 cents yield 12–13% and will be bought back with excess cash flow.
Redwood, a REIT, has cut its common dividend and says it will buy back its baby bonds using excess cash flow. Jay says the bonds trade around 80 cents on the $25 par and yield 12–13%. Because the company is actively repurchasing them, he sees an attractive, company-backed yield in the bonds.
“they're actually buying back their baby bonds. You can actually buy their baby bonds around 80¢ on the dollar, 20 on a 25 par. You know they're yielding 12, 13%”at 47:42
CoreWeave
Across the shows: 66 speakers in other episodes, 22 bullish and 70 bearish calls, past 30 days
Where everyone stands on CoreWeave
Jay Singh · Special Situations ResearchBearish
CoreWeave doesn't have a long-term reason to exist; neoclouds are commodity businesses, so he is short.
After CoreWeave rallied 25% following earnings, Jay's fund shorted it, and it also shorted a little SpaceX. His thesis is that CoreWeave, as a neocloud, has no real long-term reason to exist. The neoclouds are effectively commodity businesses. The short was partly a trade on the NVIDIA earnings lift, and he expects to eventually cover it. CoreWeave also fell 20% in 2026 after its March 2025 IPO, despite a strong year for AI.
“Core Weave, we think, doesn't really have a reason to exist in the long term... as a NeoCloud. The NeoClouds are effectively commodity businesses.”at 48:38
US stock market outlook
Where everyone stands on US stock market outlook
Jay Singh · Special Situations Research
Bear-market rallies of 5–20% reverse easily and are hard to time; this is not as bad as 2022 yet.
Jay says rallies in bear markets are very hard to time, but 5–10% rallies come easily and 20% rallies sometimes happen, and they often reverse later. He has seen this before: his firm bought stocks, sold them up 20%, and watched them round-trip in 2022. The current market is not as bad as 2022 and lacks the vicious bear-market rallies seen then. Relief would require lower inflation and a resolution to the war, which he links to Fed pricing.
“it's very hard to time rallies in bear markets, but you can see five-ten percent rallies quite easily, sometimes 20% rallies in bear markets that later reverse.”at 42:39
Jay Singh · Special Situations Research
Defensive stocks have rallied too far; he would not add more to them and is looking at beaten-down names.
Jay sees a rotation from tech into quality and defensive stocks that is stronger than anything in the past five years, even in mild recessions. Consumer staples trade at 24 times forward earnings, the highest in 27 years for that sector. Morgan Stanley says it is the biggest defensive rotation in U.S. equities in 25 years: staples, healthcare and utilities are up 13% since October 1, while the Nasdaq 100 is down 13%. He is worried about adding to these crowded defensive names and says his firm is looking at beaten-down names that are not correlated with broader tech.
“So I would actually be worried about adding more and more to some of these defensive names that are rallied a lot”at 44:46
US–China AI & chip race
Where everyone stands on US–China AI & chip race
Jay Singh · Special Situations Research
China still wants Nvidia GPUs, but its domestic chips and memory lag for complex models.
Singh says China absolutely would still have demand for GPUs. Its AI industry keeps growing despite the sell-off in AI stocks. He argues that China's domestic GPUs and memory chips are not yet advanced enough for the most complex models, so Nvidia's chips still have a role. He agrees with the critics that Nvidia may be earning revenue from China through intermediaries, and the researcher short seller who said so was early to the point.
“So China absolutely, would have the demand for GPUs. ... their AI industry continues to grow and their GPUs and even you know their memory chips are not as advanced as they need to be for the more complex models.”at 19:49
Cybersecurity
Where everyone stands on Cybersecurity
Jay Singh · Special Situations Research
Cybersecurity companies look very interesting after this sell-off, as AI makes cyber threats harder to contain.
A run of outages has hit almost every day: a Microsoft 365 outage on November 12 (the longest Outlook outage on record), an FAA nationwide ground stop on November 13 that cancelled 20,000 flights, a United Airlines outage on November 16, bank ATM failures, a Schwab outage on November 17, and NSA classified networks and Apple services down. Singh says Iran could be behind some of the attacks, and that hacking is getting easier as AI becomes more powerful. A lawsuit against OpenAI over inadequate hacking safeguards supports this. He sees cybersecurity companies as attractive because of this threat backdrop.
“So you know AI will continue to be very challenging and that's one reason why we think the cybersecurity companies are very interesting in this sell off.”at 28:48
IPOs, M&A & deal activity
Where everyone stands on IPOs, M&A & deal activity
Jay Singh · Special Situations Research
Recent 2026 tech IPOs pop on demand, then drop as lockup unlocks approach; avoid buying early.
Jay describes a 'pop and drop' pattern in recent mega IPOs. Small floats and heavy first-day demand push prices up, then investors price in upcoming unlocks and sell until positive earnings arrive. Nearly all major 2026 tech listings have retreated sharply after debut, and on average 2026 second-day IPOs trade 23% below first-day prices. He cites Cerebras and SpaceX as examples. He contrasts this with Facebook, whose IPO unlock fears faded as earnings grew into the valuation. His practical takeaway: without an allocation, be careful buying on the first day and do not buy on the way up without trimming. He tends to avoid IPOs and buys after the six-month unlock, because selling pressure is heavy and the timing is unpredictable.
“they've followed, what we call the pop and drop pattern, where they pop initially on huge demand, and then when people start to factor in the upcoming unlocks, they sell off until you get positive earnings news.”at 38:36
Market sentiment, positioning & volatility
Where everyone stands on Market sentiment, positioning & volatility
Jay Singh · Special Situations Research
Frequent 2% S&P 500 daily moves now exceed the 2010s and 2020, matching only the 2000 tech crash.
Citing Jefferies, Jay says the number of days with an absolute S&P 500 move above 2% is running higher than in 2020 and the 2010s. The only comparable period is 2000, during the tech crash. Combined with institutions cutting tech allocations at the fastest pace since June 2006 (per BofA), he reads the tape as a stressed, defensive market rather than a normal correction.
“the number of days, you know, with greater than an absolute S and P 500 move of 2% according to Jefferies is on a run rate now higher than the 2020 period, higher than the 2010s. The only period where we've seen this many 2% absolute S and P 500 moves, is in 2000 during the tech crash.”at 43:49
US rates & the Fed
Where everyone stands on US rates & the Fed
Jay Singh · Special Situations Research
He is very confident the Fed hikes in December; a drop to two priced hikes would spark a relief rally.
Jay expects the Fed to hike in December and is very confident of this. Markets are pricing about 3.5 hikes. If inflation data starts to come down and the war is resolved, he thinks pricing could fall to two hikes, which would bring a relief rally in December. Right now, with war uncertainty, the market lacks that catalyst. He also says the company rate-hike outlook is part of why closed-end funds and preferreds, with high yields, look attractive.
“I think after the Fed hikes in December, very confident that they're gonna hike. If the inflation data starts to come down and we get a resolution to the war”at 49:22
Across the shows
What other episodes Chorus covered said about the same companies in the 30 days up to this one.
- NVIDIA178 speakers · 182 bullish, 182 bearish calls
- Anthropic153 speakers · 269 bullish, 133 bearish calls
- CoreWeave66 speakers · 22 bullish, 70 bearish calls
- OpenAI42 speakers · 20 bullish, 33 bearish calls
- Cerebras Systems5 speakers · 3 bullish, 1 bearish calls
Topics
- AI-enabled cyber threats and cybersecurity stocks
- Anthropic IPO financials
- China demand for AI chips
- IPO aftermarket performance
- IPO pricing and post-listing performance
- Jensen Huang departure and leadership
- Nvidia forward valuation multiple
- aftermarket aircraft parts
- baby bond buybacks
- balance sheet capacity to back guarantees
- bear market rallies
- cash burn and SoftBank funding
- defensive sector rotation
- digital advertising growth
- export restrictions impact on Nvidia sales
- gross margin trajectory
- guidance and growth outlook
- index volatility
- neocloud business model
- neocloud financing guarantees
- open-source competition
- rate hikes
- safety and governance risks
More from The David Lin Report
- Anthropic IPO: Are Insiders Already Dumping Shares? | Will Rhind · November 20, 2026
- 50% Market Crash In 2027: Investor Called Top, Now Has Dire Warning | Felix Zulauf · November 18, 2026
- Tech Stocks Face 70% Crash: What Happens To Economy? | Steve Hanke · November 17, 2026
- Markets ‘Spiraling Out Of Control’: When Will Stocks Bottom? | Chris Vermeulen · November 16, 2026
- Worst Housing Demand Ever: Where Prices Fall In 2027 | Reventure CEO · November 16, 2026
Ask your AI about this episode
Chorus gives Claude, ChatGPT and other assistants every view from this episode and every other show Chorus follows. Ask who agrees, who changed their mind, and what else was said about the same companies.
Chorus reports what speakers said. It isn't investment advice. How Chorus works · About Chorus