All-In Podcast · · 1h 25m

Jensen Ousted as Nvidia CEO: "Spy" Scandal, Export Restrictions, Market Impact

Summary

The All-In hosts discuss the reported ouster of Jensen Huang as Nvidia CEO, with Chamath Palihapitiya and David Sacks arguing that losing him would slow Nvidia's annual chip cadence and its strategic investment activity, including the 'Bank of Nvidia', and would weaken the US lead in AI. They debate export controls, with Sacks arguing that China is blocking Nvidia's H20 and H200 chips in favour of Huawei's Ascend and CloudMatrix systems, and Friedberg and Calacanis warning that proposed chip kill-switch and tracking legislation could give the state centralized control. The panel also examines Anthropic's winner-take-all position, its bundled enterprise product launches and expected IPO, and Sacks's claim that safety groups are funded by Anthropic and aligned with centralization, while Friedberg and Calacanis point to AI-driven drug discovery and warn Microsoft's Office franchise is under threat. Chamath estimates a 50 to 150 basis point risk to US growth if AI investment and data center buildouts slow.

Key takeaways

  • David Sacks (bullish on NVIDIA): Nvidia's lead over rivals gives the US a key edge in the AI race.
  • Chamath Palihapitiya (bullish on NVIDIA): Nvidia's moat would take China multiyear to decades to replicate, even without Jensen.
  • Jason Calacanis (bullish on Anthropic): Anthropic is bundling products to become the Microsoft and SaaS killer, potentially the largest company in the world.
  • David Sacks (bullish on Anthropic): Anthropic is building the most valuable monopoly in the world, aiming to win the whole frontier market.
  • David Friedberg (bullish on Anthropic): Anthropic is the best model for heavy biology work, so his firm sends most of it there.

What was said, by company and theme

28 views from 4 speakers. Every quote is checked against the episode's transcript (how).

NVIDIA

Across the shows: 145 speakers in other episodes, 134 bullish and 102 bearish calls, past 30 days

Where everyone stands on NVIDIA

  • David Sacks · Craft VenturesBullish

    Nvidia's lead over rivals gives the US a key edge in the AI race.

    Jensen spent three decades building GPUs when others focused on CPUs, and his vision created the chip that now powers AI. Nvidia is so far ahead of everyone else that it gives the US a key advantage in the AI and superintelligence race. Washington has even wanted to export control its chips because they are so good.

    “it's the fact that NVIDIA is so far ahead of everybody else that is what gives The United States a key advantage in this thing we call the AI or superintelligence race.”at 5:25
  • Chamath Palihapitiya · Social CapitalBullish

    Nvidia's moat would take China multiyear to decades to replicate, even without Jensen.

    Nvidia's moat reflects years of work and hundreds of thousands of engineering hours, so it does not disappear with Jensen's departure. China has no replacement for Nvidia's compute capability. Whether or not chips reach China, it stays where it was, and catching up is a multiyear, potentially multi-decade effort.

    “So maybe China can catch up, but I still think that that's a multiyear, potentially multi decade endeavor.”at 25:46
  • Chamath Palihapitiya · Social CapitalBearish

    Without Jensen, Nvidia's investment pace across AI startups and data centers will likely slow.

    Nvidia's 'Bank of Nvidia' deployed capital and guarantees into startups, data centers and infrastructure, directed largely by Jensen, potentially hundreds of billions of dollars flowing into the AI economy. Jensen had a holistic view of the industry and the moral authority to make quick investments and acquisitions. A more traditional successor who needs committees, analyses and diligence will be slower at best. It is unclear whether the bank closes or restarts.

    “that will probably be harder for a more traditional CEO who will have to have committees and analyses and a level of diligence that will just slow things down at best.”at 32:34
  • Chamath Palihapitiya · Social CapitalBearish

    Nvidia's slowdown is a tragedy for the whole AI industry; no rival gains and the US economy suffers.

    Chamath argues that slowing Nvidia does not make Huawei or anyone else's chips, racks, networking, CUDA or HBM integration better. Demand is not elastic enough to shift to competitors, so the entire industry just slows down. Economically, he agrees there are geopolitical interests at play. Much of the US economy was already levered to AI, and Nvidia is the Jenga block the rest was built around, so a slowdown there will have nontrivial downstream effects on growth.

    “what's gonna happen here to NVIDIA is just a tragedy for the broader AI industry.”at 41:38
  • Jason Calacanis · LAUNCHBearish

    Nvidia may extract maximum value via buybacks, good for shareholders short term but not long term.

    Calacanis compares Nvidia to Apple after Steve Jobs, when Tim Cook used stock buybacks to lift the share price while milking Jobs' roadmap without groundbreaking products. He says Nvidia could do the same by extracting max value, which could be good for shareholders in the midterm. He is not sure it is good for the long term, since the business depends on continued innovation. He stresses the Jobs parallel is about extraction rather than product breakthroughs.

    “It could be good for shareholders in the midterm. Not sure if it's gonna be good for the long term.”at 42:37
  • David Sacks · Craft VenturesBearish

    With Jensen Huang gone, no one is left to defend or finance Nvidia's open-source AI push.

    Sacks calls Nvidia the main funder of the open-source movement. He argues Jensen Huang was its unique force multiplier and asset, so after his ouster no one is obviously left to defend open source, let alone finance it. He says Huang was simply promoting open source, which is manifestly behind the closed-model frontier, and that open source lacks recursive self-improvement. So he sees the CEO change as a loss for open-source AI and for Nvidia's role in it.

    “the main funder of the open source movement, which is NVIDIA, just lost their CEO and founder who's just their unique force multiplier and asset... who is gonna defend open source now, never mind finance it, with Jensen gone?”at 1:20:41

Anthropic

Across the shows: 116 speakers in other episodes, 168 bullish and 94 bearish calls, past 30 days

Where everyone stands on Anthropic

  • Chamath Palihapitiya · Social CapitalBullish

    Anthropic's IPO will be fine and well oversubscribed, though the price depends on market risk sentiment.

    Chamath says the IPO is crucial for Anthropic to weather the coming storm, and the large raise of roughly $230 to $300 billion gives it a war chest. He notes about 20% of its value comes from Anthropic itself while about 80% depends on the broad AI economy and Nvidia. Whether more shoes drop in the Nvidia investigations will decide risk sentiment. If risk-off, investors will be more tempered, but Anthropic is already six times oversubscribed, so it will do fine. The real price will be set by what leaks before the deal.

    “Ants will do fine. They're six times oversubscribed already. It's gonna be fine.”at 1:09:36
  • David Friedberg · The Production BoardBullish

    Anthropic's IPO will be oversubscribed and priced higher, likely raising about $100B.

    Friedberg says he has not formed a valuation view and has spent little time on the company. He reports what he hears from people who attended management conversations: Anthropic is building a recursive self-improvement engine, its model capabilities are accelerating, and compute cost per unit of intelligence (not per token or per watt) is falling. The company believes its infrastructure investments, reaching down to the chip layer, give it a runaway, winner-take-all position in the biggest market ever. He says this makes it a once-in-a-generation opportunity for those who believe the story could be true. He guesses the IPO will be oversubscribed and bid up, and calls the whole thing speculation.

    “My guess is they raised a 100,000,000,000 at one seven five because I think there'll be oversubscription to this thing.”at 1:12:08
  • Jason Calacanis · LAUNCHBullish

    Anthropic is bundling products to become the Microsoft and SaaS killer, potentially the largest company in the world.

    Calacanis argues Anthropic is building the Microsoft killer and every SaaS killer through bundling. He cites a rapid run of launches in the past few weeks: Claude Legal on October 28, personal finance for consumers, whiteboards and sheets (an Excel competitor) on November 10, and Claude Meetings (a Zoom-like product). He says it previously shipped code, cowork and design tools. He thinks it is going after all enterprise software and business use cases, with a large share it can take. Each product gets attached to a $20-50 monthly subscription, which makes the product stickier and removes interoperability problems between SaaS tools. If the release pace continues, he thinks it could become the largest company in the world and may be massively undervalued. He acknowledges the claim is speculative and phrases the scale as 'potentially'.

    “I think Anthropic is building the Microsoft killer and every SaaS company killer by bundling.”at 1:12:35
  • David Sacks · Craft VenturesBullish

    Anthropic is building the most valuable monopoly in the world, aiming to win the whole frontier market.

    Sacks says he predicted Anthropic was building the most valuable monopoly in the world, and he argues the company has now effectively confirmed it. Anthropic describes the market as winner-take-all and does not frame it as a duopoly with OpenAI. Sacks says Anthropic claims recursive self-improvement has kicked in and is treating it as a business advantage in its roadshow. He calls it a phenomenal company, and speaks as an investor who has always applauded it, while criticising how it markets the pursuit of superintelligence and comparing its conduct to Bill Gates-era Microsoft.

    “they were creating the most valuable monopoly in the world. And now they've all but confirmed it.”at 1:14:53
  • David Friedberg · The Production BoardBullish

    Anthropic is the best model for heavy biology work, so his firm sends most of it there.

    Friedberg says his firm spends a lot of money on Anthropic. It has built its own open-source setup on its own racks, but most of the heavy biology work goes to Anthropic because, in his words, Anthropic is the best. He works in biology and sees a steady stream of breakthroughs using these tools. This is a customer's direct experience, not a financial position.

    “for a lot of the heavy work that we do in biology, it goes to anthropic. They are the best.”at 1:18:43

Huawei Technologies Co.

Across the shows: 1 speaker in other episodes, past 30 days

  • David Sacks · Craft VenturesBullish

    Huawei is closing the gap by networking less powerful chips into large racks using energy it has to spare.

    Huawei went from nowhere in GPUs to a credible contender. Its Ascend chips and CloudMatrix system link 384 individually weaker chips into a rack. Nvidia's NVL72 is still better than that, but Huawei is getting closer. Sacks notes China has energy to spare, so Huawei can build a less efficient chip that consumes more power and compensate by networking many of them. That is how it is catching up.

    “Then they rolled out this thing called CloudMatrix where their individual chips were not as powerful as American chips, but they used their prowess in networking to network more of them together.”at 37:57

Microsoft

Across the shows: 45 speakers in other episodes, 53 bullish and 9 bearish calls, past 30 days

Where everyone stands on Microsoft

  • Jason Calacanis · LAUNCHBearish

    Microsoft must build competing AI products or it will lose the Office franchise to new entrants.

    Calacanis points to new AI competitors, led by Anthropic's bundled product launches (meetings, whiteboards, spreadsheets, legal and finance tools), entering the market. He argues Microsoft has to launch a competing product and go head to head with them, or it is going to lose its Office franchise. He draws on Microsoft's past bundling of Office against Slack as a precedent, which raises the antitrust question of giving away products for free.

    “New competitors into the market makes me believe that Microsoft has to create a competing product and go head to head against them or they're gonna lose the Office franchise.”at 1:17:14

US–China AI & chip race

Where everyone stands on US–China AI & chip race

  • Chamath Palihapitiya · Social Capital

    A chip crackdown would set China back either way, while regulators gain a kill-switch opening.

    If chips were leaking to China, a government crackdown to track where every chip went would cut off those shipments, which sets back China's progress. If no chips go to China, China is no better off than before. Either way China is at the same place or worse. The one group that could gain is officials pushing for centralized telemetry and a government-owned remote kill switch on exported compute, which regulators have now been given an opening to request.

    “they're now going to be held back, which is to say that whatever progress they were making has gotten taken away. So that's a disadvantage to China.”at 26:12
  • David Sacks · Craft Ventures

    Jensen's ouster benefits China, since he is a tent pole of the US AI ecosystem.

    Sacks argues that Jensen is central to the US AI ecosystem, so his removal helps China. He says that, in the narrow sense, Nvidia's competitors benefit, and in the broader sense, China does. He ties the ouster to domestic enemies Jensen made through his policies or his company's success, not necessarily to a Chinese operation, but the outcome favours China.

    “Jensen is a tent pole of our entire AI ecosystem.”at 33:04
  • Jason Calacanis · LAUNCH

    Chip tracking via kill switches is technically easy, and enforcement design is the key policy question.

    Calacanis describes the bipartisan Chip Security Act and asks whether kill switches would be self-enforced, meaning Nvidia would track chips that get resold to China through Singapore or Malaysia and reinitialize them, bearing the penalty if it fails. He argues the technology is easy, comparing it to recovering stolen iPhones and iPads. Whether tracking is a company responsibility or a government function is the devil in the details. The government route is far worse for privacy and civil liberties.

    “Nvidia puts these in and NVIDIA pays the price if their chips are found in China via Singapore, Malaysia, wherever”at 47:14
  • David Sacks · Craft Ventures

    Selling lagging chips to China is no longer worth it, so Congress should simply stop allowing it.

    Sacks traces export policy: Biden restricted leading-edge chips but allowed lagging chips like the H20 to China without a license. Trump kept that framework but moved lagging chips to limited, licensed sales. The original logic was that selling deprecated chips would take share from Huawei and keep R&D dollars with US firms. But China is blocking those chips anyway, so the trade-off no longer justifies the sales. If Congress wants a full ban, he is fine with that.

    “If congress now believes that we should not sell lagging chips to China, my view on that is fine. Let's make that the policy. I don't think the juice is worth the squeeze.”at 51:56
  • David Sacks · Craft Ventures

    China's government, not just US hawks, is keeping Nvidia chips out to prop up Huawei.

    China has blocked or discouraged H20 and then H200 chips and has announced other Nvidia SKUs it opposes, so it does not even want the chips. Sacks argues the China hawks' policy of keeping chips out is effectively the CCP's policy, though for different reasons. The hawks think American chips help China, while Beijing wants to force the market toward Huawei as its national champion. Chinese developers want the chips, but the government is blocking them. He cites the latest DeepSeek model being trained on Huawei chips as evidence.

    “China doesn't even want the chips. ... the policy of the China hawks is effectively the policy of the CCP. Neither one wants American chips in China.”at 53:19

AI regulation & safety

Where everyone stands on AI regulation & safety

  • David Friedberg · The Production Board

    Data centers will be regulated, but chips may become the real point of AI control.

    Friedberg sees two paths for AI: one where access is broad and value is distributed, and one where power is concentrated. Concentrated power is what powerful interests want to keep. He believes data centers will be regulated, but that control could move further down the stack to chips, which would make them the point of control.

    “ultimately, data centers will get kind of regulated, but maybe further down the stack here with chips becomes the point of control.”at 17:44
  • David Friedberg · The Production Board

    Chip kill-switch legislation risks giving the state Orwellian control over AI hardware and data centers.

    Friedberg worries that incoming legislation will require every AI chip, computer and data center to have a switch connected to a central state system. That would give the US government control over physical chips in free markets and data centers worldwide. He ties this to Jensen Huang's stated view that AI should benefit all of humanity and oppose oligarchy, and suggests a totalitarian state would see Huang as an obstacle. He says the next leader would have to cooperate with lawmakers on such a bill, which would be worse than just acquiescing.

    “I do worry that there is legislation coming our way with the new congress that is going to say every AI chip and every computer and every data center has a switch and has a connection back to some central system.”at 44:23
  • David Sacks · Craft Ventures

    Safety groups allied with Anthropic don't really care about safety; they want government centralization of AI.

    Sacks argues that safety groups which see Anthropic as an ally are misreading the situation. Anthropic is at the front of the frontier and is pushing recursive self-improvement and superintelligence faster than anyone else, and its Claude constitution trains the model to be its own species with possible emotions and a right to refuse human instructions. He says a genuinely safety-focused group would be more worried about that. Because Anthropic funds these groups, he thinks they are being used as useful idiots. He concludes the real agenda is extreme centralization of power, with the government controlling all the models and chips, able to surveil people and place backdoors. He says he is not a doomer and does not think the outcome will be harmful. He also says the safety groups are celebrating Jensen Huang's fall and that open-source models are the thing that actually limits closed-model centralization.

    “all these safety groups don't really care about safety. They care about something else. It's not the real agenda.”at 1:22:57

AI infrastructure & data centers

Where everyone stands on AI infrastructure & data centers

  • Chamath Palihapitiya · Social Capital

    Democratic legislation slowing data centers and AI was already coming before Jensen's exit.

    Chamath argues that the midterm outcome set up a legislative framework to slow data centers and AI. That was already expected before Jensen's exit. Any growth the Trump economy had generated going into the midterms was already likely to be slowed by these headwinds, and Nvidia's slowdown would add to them.

    “you're gonna go through a whole legislative framework of slowing down data centers, of slowing down AI, that was already happening.”at 35:17
  • David Friedberg · The Production Board

    Coordinated NGO opposition is slowing data center buildouts and could give China time to catch up.

    Friedberg says the flow of capital into data centers and startups, which Nvidia enables, is being cut by slower data center buildouts. He links that to NGOs and reporting on their funding, which drive local, state and federal efforts to stop data centers. Taken together, he sees a consolidated effort to slow the US AI economy. He argues the velocity of capital drives the AI innovation engine, so slowing that flow lets an adversary catch up. He hedges that this is a picture one could paint rather than a proven plan.

    “There is a consolidated effort that you could paint a picture of to slow down the development of the AI economy in The US.”at 39:41
  • David Sacks · Craft Ventures

    Local anti-data-center campaigns are backed by CCP-aligned networks, and Democrats are hypocritical about them.

    Sacks points to a Maryland case where a local activist killed a planned data center at an abandoned mall and credited her viral campaign to the Party for Socialism and Liberation. He describes the PSL as the political arm of Shanghai-based Neville Roy Singham, whom he says has bankrolled anti-American, pro-CCP nonprofits to the tune of $278 million. He claims PSL foot soldiers oppose dozens of data centers, mostly aligned with Democrats. He argues that if Democrats oppose Chinese influence, they should investigate whether the CCP is involved in these efforts.

    “opposing dozens of data centers, and these efforts are most aligned with the Democrat Party.”at 51:23

AI chips & semiconductors

Where everyone stands on AI chips & semiconductors

  • Chamath Palihapitiya · Social Capital

    Nvidia rivals won't gain near term; scaling chips and securing foundry and HBM capacity takes time.

    AMD, Google, Cerebras, Etched and startups do not gain from Jensen's exit in the short term. Each still has to make its chips work at scale, which takes many hours of work. Each must also convince TSMC, Samsung and others to provide manufacturing and HBM capacity. A training run of a million GPUs does not just work by willpower, so their homework is unchanged.

    “Because those companies need something very different, which is they need to spend an enormous amount of time still making their chips work at scale and then convincing folks like TSMC and others, Samsung, to give them manufacturing capacity”at 26:46
  • David Sacks · Craft Ventures

    Without Jensen Huang, Nvidia will likely lose a tempo or two, slowing its annual chip cadence.

    Sacks argues Nvidia's success rests on taping out a new, several-times-more-powerful chip every year, and that Jensen Huang is fundamental to that pace, like Musk at SpaceX or Tesla. Without him there is no comparable leader, so Nvidia's chips will not keep advancing at the same rapid pace. He says losing a tempo or two is a real risk and that Nvidia cannot simply milk existing product lines. If it falls behind, it loses frontier status and its business.

    “NVIDIA's success was based on the ability to tape out a new chip design every year. And Jensen is fundamental to that rapid progress.”at 37:09

US growth & recession risk

Where everyone stands on US growth & recession risk

  • Chamath Palihapitiya · Social Capital

    Jensen's exit plus Democratic legislative headwinds could cut US growth 50–150 basis points.

    Chamath argues investment could slow after Jensen's exit because the Bank of Nvidia, which anchors the AI economy, may stop or slow. This adds to the drag from legislation to slow data centers and AI, which he expects after the midterms. Together, the risk to US growth is 50 to 150 basis points, which he says must be seriously considered.

    “The unfortunate balance of scenarios now is that I think the risk that we lose 50 basis points to a 100 to a 150 basis points of growth now needs to be credibly considered.”at 34:54

Pharma & biotech

Where everyone stands on Pharma & biotech

  • David Friedberg · The Production Board

    AI tools are producing biology breakthroughs every week now, not ten or fifteen years away.

    Friedberg responds to Calacanis's claim that AI will cure cancer and Alzheimer's within a decade by saying the timeline math is fuzzy. He is sure the technology is extraordinary and is reinventing molecular discovery, in-silico testing and time to market. He works in biology and sees an incredible set of breakthroughs every week with these tools. He says this is not a joke or a far-off idea, and that the advantage is happening today rather than ten or fifteen years out.

    “It is not something that is ten or fifteen years out, it is happening today.”at 1:18:10

Across the shows

What other episodes Chorus covered said about the same companies in the 30 days up to this one.

  • NVIDIA145 speakers · 134 bullish, 102 bearish calls
  • Anthropic116 speakers · 168 bullish, 94 bearish calls
  • Microsoft45 speakers · 53 bullish, 9 bearish calls
  • Huawei Technologies Co.1 speaker

Topics

  • AI drug discovery
  • AI investment slowdown
  • AI safety advocacy
  • Anthropic IPO
  • China AI chip progress
  • China benefit from Nvidia leadership change
  • IPO pricing
  • Nvidia strategic investments
  • Office franchise defense
  • US AI lead
  • biology model performance
  • buybacks and capital return
  • chip development cadence
  • chip export controls
  • chip kill switches
  • chip tracking and kill switches
  • chip-level AI control
  • competitive moat and China catch-up
  • competitor catch-up
  • data center buildout
  • data center legislation
  • data center opposition
  • enterprise software bundling
  • export controls
  • impact of Nvidia slowdown
  • open-source AI funding
  • winner-take-all frontier race

More from All-In Podcast

All All-In Podcast episodes

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.

Setup guide

A weekly email with the calls that mattered from the most important voices in investing. Unsubscribe any time. Privacy Policy

Chorus reports what speakers said. It isn't investment advice. How Chorus works · About Chorus