<h3>Key takeaways</h3>
<ul>
<li>Jensen Huang resigned as Nvidia CEO hours after reports of a second affair, this time with a direct report, Kimberly Powell.</li>
<li>CFO Colette Kress becomes interim CEO; Nvidia has shed more than $600 billion in market value this week.</li>
<li>Short seller Christian Lamarco of Culper Research, who flagged Nvidia's "China problem" six months ago, saw his thesis go viral.</li>
<li>Dylan Patel names Charlie Kawas and Johnny Srouji as possible successors; Shaan Puri instead floats Nikesh Arora.</li>
<li>Hosts debate whether China actually wanted Jensen out, with Jordan Schneider and others questioning the theory.</li>
</ul>

<h3>Jensen Huang is out after a second affair allegation</h3>
<p>The hosts open with the news that Jensen Huang resigned as Nvidia CEO less than 24 hours after reports surfaced of a second extramarital affair, this one with a colleague who reported directly to him, Nvidia's VP of healthcare Kimberly Powell. The New York Times reported the two were seen being affectionate at a San Jose hotel during GTC in March. The host calls it "the greatest business tragedy of my lifetime," while noting the affair with a subordinate is a clear violation that forced the board's hand: "It's not a complex morality debate... It's cut and dry."</p>
<p>The hosts note Nvidia's board had no succession plan and no chair until naming longtime director Harvey C. Jones to that role just before the announcement. CFO Colette Kress, at Nvidia since 2013, becomes interim CEO. The hosts say her background gives them confidence in "business as usual" on operations, though they flag that she's been framed as interim rather than a clear successor.</p>

<h3>Can Nvidia still be the "central bank of AI" without Jensen?</h3>
<p>A central debate is whether Nvidia's model of using its balance sheet to bankroll AI customers can continue without Jensen's personal dealmaking. One host argues Kress can clearly manage the "make GPU, sell GPU" side, but questions whether talent recruiting suffers without a founder-mode CEO able to offer the pitch of joining the biggest, fastest-growing player. The hosts also flag TSMC's relationship with Nvidia as resting on "mutual respect," and worry rivals building custom chips (TPUs, Trainium, AMD) could take advantage.</p>
<p>They also raise the possibility that capital allocation shifts "from Jensen acting as the banker... to investment by committee," which one host says "is much, that's just going to be a lot slower."</p>

<h3>The short seller who called Nvidia's China problem</h3>
<p>The hosts recap a Wall Street Journal story on Christian Lamarco of Culper Research, who published a report six months ago arguing Nvidia has "a significant China problem," tied to an alleged close relationship between Jensen and an executive named Alice Wong at a Southeast Asian business partner. Culper wrote it was "not a bet against AI." The report was largely ignored until this week, when Nvidia's stock fell and the firm's posts drew millions of views.</p>
<blockquote>"We are short Nvidia for one reason. The company has a significant China problem." — Culper Research</blockquote>
<p>The hosts note Nvidia has lost "more than 1 trillion in market value" and is down roughly 20% since the report, though it's unclear whether Lamarco profited, as he declined to disclose his position or earnings.</p>

<h3>Reactions split over succession and whether Jensen should have stayed</h3>
<p>Social media reactions quoted on the show diverge sharply. Analyst Dylan Patel suggested the board "needs to chase down Charlie Kawas and Johnny Srouji" as the only viable replacements. Investor Shaan Puri instead proposed Nikesh Arora, citing his record at Palo Alto Networks. Others, like Doug Ollafin, argued Jensen should have stayed: "He stays. Make a transition plan... The company's not set up to run without him."</p>
<blockquote>"You can't have an affair with a direct report. It's over." — Tae Kim, as quoted by the host</blockquote>
<p>Commentator Bern Hobart is quoted noting Jensen "lives in a state where you can't enforce non-competes," fueling host speculation about a possible comeback, though they say it would be harder than a Sam Altman-style return given active congressional and DOJ scrutiny.</p>

<h3>Did China want Jensen gone? Hosts are unconvinced</h3>
<p>The hosts discuss theories, including one from Nick Carter suggesting a "deep state coup in conjunction with China Hawks and the admin," and a claim attributed to Shaan Puri that "China wants Jensen out." Jordan Schneider is quoted pushing back, asking what exactly was on China's wish list that Jensen didn't deliver. The hosts say they can't verify the claim and note China's own restrictions on Nvidia GPU imports complicate any simple narrative.</p>
<p>They also read a long post from Gavin Baker arguing that "well-meaning, patriotic China hawks" who pushed export restrictions may have pushed China toward developing its own AI hardware, which he warns could make the world "much less stable" if China catches up and shares the technology with Russia, Iran, and North Korea. One host says he "totally disagrees" with Baker's closing point but doesn't elaborate further.</p>

Jensen Huang resigns as Nvidia CEO after second affair scandal, Kress named interim chief

TBPN's hosts break down Nvidia's leadership crisis, the short seller who called it, and what losing Jensen means for the "central bank of AI."

Jensen Huang Resigns as NVIDIA CEO | Diet TBPN

<h3>Key takeaways</h3>
<ul>
<li>Dylan Patel says Nvidia deals that relied on Jensen Huang's personal authority will now "go to committee," slowing decisions.</li>
<li>Patel names Ian Buck and Jonah Alvin as internal succession candidates, but prefers external picks Charlie Kawwas of Broadcom and Johnny Srouji of Apple.</li>
<li>Patel argues open-source AI is the biggest loser, since Nvidia-backed labs like Reflection and Thinking Machines rely on open models tied to China.</li>
<li>He says Anthropic benefits versus OpenAI since OpenAI is "90 plus% Nvidia" while Anthropic is "one-third or less Nvidia."</li>
<li>Patel predicts venture capital is now "out of the lab trade" as OpenAI and Anthropic move toward public liquidity, leaving VCs reliant on Jensen's backing of inference and application startups.</li>
</ul>

<h3>Nvidia's deal pipeline loses its decision-maker</h3>
<p>Dylan Patel argues that Jensen Huang's exit removes the single figure who could unilaterally commit Nvidia's capital and supply to AI companies. He says deals that used to rely on Huang personally promising large investments will now face slower, more bureaucratic approval. "Now it's like a committee of people doing stuff and that's just going to make decisions slow," he says, adding that Nvidia could technically fall behind as a result.</p>
<p>Patel breaks the risk into three parts: no clear technical successor, a supply chain that responded to Huang's personal charisma and dealmaking rather than cold business logic, and an ecosystem of open-source labs and application companies that Nvidia has financially propped up. He doubts a new CEO could convince suppliers to hand over "60% of my capacity or 70% of my capacity... at a price that's below market rate" the way Huang could.</p>

<h3>Possible successors: internal picks weighed against outside options</h3>
<p>Internally, Patel names Ian Buck and Jonah Alvin as the two most plausible successors, noting Alvin is "technically like the closest to the leader" of Nvidia's technical organizations, while Buck is more charismatic and better suited to dealmaking. He is skeptical either can match Huang across both technical and supply-chain dimensions simultaneously.</p>
<p>His preferred external candidates are Charlie Kawwas, described as Broadcom's long-serving number two, and Johnny Srouji, Apple's head of hardware. He dismisses other names floated, including AWS's Andy Jassy and Google's chip leadership, arguing none combine the technical depth and supply-chain instincts Nvidia's position requires.</p>

<h3>Open-source AI and Chinese labs face a reckoning</h3>
<p>Patel says Chinese AI labs have worked closely with Nvidia even without direct chip sales, and that labs reliant on Nvidia, such as those using Huawei or diversified chips, are now adjusting their assumptions. He singles out Zhipu and DeepSeek as already diversifying across non-Nvidia accelerators.</p>
<blockquote>"I think the biggest loser is the open source ecosystem for sure." — Dylan Patel</blockquote>
<p>He argues that with Huang's credibility "out the window" amid investigations, defenders of looser China chip policy have lost influence, likely leading to tighter restrictions. He contrasts OpenAI, which he says is "90 plus% Nvidia," with Anthropic, "one-third or less Nvidia," arguing the fallout benefits Anthropic over OpenAI since Anthropic pushes open-source weaker, closed-source winning.</p>

<h3>Venture capital loses its biggest ally</h3>
<p>Patel says venture capital is now largely "out of the lab trade" as OpenAI and Anthropic pursue public-market liquidity, calling OpenAI's last funding round "not that great." He warns the bigger risk is downstream: inference clouds and application companies that depended on Nvidia-backed financing and compute access.</p>
<blockquote>"Jensen is the venture ecosystem's strongest ally and he's propped up a lot of companies." — Dylan Patel</blockquote>
<p>He questions whether firms can scale from "a billion dollars of revenue to 10 billion" without secure compute financing, and whether application companies like Harvey and Sierra will be forced onto closed-source models, becoming "wrappers" rather than owning their customer relationships.</p>

<h3>AI 2027 predictions and the China "wake-up" question</h3>
<p>Discussing the AI 2027 forecast, Patel calls it "a big stretch" but notes roughly 70% of its verifiable predictions this year "have turned out to be accurate." He remains skeptical about diffusion-related predictions but says reinforcement learning self-improvement "is looking more and more true."</p>
<p>On whether China is "waking up" to AI competition, as the forecast anticipated, Patel is uncertain, noting China could simply tighten its own supply chain leverage given every Nvidia and Google server contains Chinese content. He also pushes back on the idea that the Chinese businesswoman linked to Huang was a deliberate honeypot, calling her instead "a well-connected individual" likely passing information under pressure rather than acting as a classic spy.</p>

Dylan Patel: Jensen Huang scandal could hand control of AI's "central bank" to a committee

Semianalysis's Dylan Patel argues Jensen Huang's exit would slow Nvidia's dealmaking, hurt open-source AI, and strip venture capital of its biggest backer.

&quot;Jensen Huang&#39;s Exit Could Change the AI Industry Forever&quot; - Dylan Patel

<h3>Key takeaways</h3>
<ul>
<li>Taiki Maeda says he was forced to sell his leveraged Zcash perps position on Tuesday, banking somewhere around $2.3 million in profit after being up $6.4 million at the peak.</li>
<li>He argues Zcash's real catalyst is the Tachyon upgrade, due around late January, which targets 50,000 TPS and post-quantum security, not just the privacy narrative.</li>
<li>Taiki sets a personal target of ZEC above $2,000 next year and floats $20,000 if it ever seriously threatens to flip Bitcoin.</li>
<li>He says Bitcoin could "trade like Ethereum last cycle," ranging for years amid a quantum and scalability "crisis of confidence."</li>
<li>He flags PRL, Monero and Zama as other assets showing similar privacy/alternative-store-of-value accumulation signals, though he holds none of them.</li>
</ul>

<h3>A forced liquidation closes out a leveraged Zcash trade</h3>
<p>Taiki Maeda opens by explaining that he had his "entire net worth" in Zcash perps on leverage through Variational, and that a sharp market dump this week forced him to sell despite remaining bullish. He says the position's total P&L, including remnants of an earlier ETH short, was around <strong>$3 million</strong>, with the Zcash-specific gain closer to "<strong>two and a half or 2.3</strong>" million, down from a peak of <strong>$6.4 million</strong>.</p>
<p>He describes the breakdown: Zcash had been ranging between roughly $1,100 and $1,400 before breaking down, prompting him to sell "coping" along the way. He says his spot bags in Zcash, Hype and Bitcoin remain, and that his portfolio is now mostly cash with no leverage, waiting for a Variational airdrop next month as a possible next trade.</p>
<blockquote>"I'd rather be up a lot, not sell, and then roundtrip some, and close at a healthy profit, than to not risk anything at the beginning, have less drawdowns, but be up a couple hundred thousand." — Taiki Maeda</blockquote>

<h3>The case for ZEC above $2,000: Tachyon, not privacy</h3>
<p>Taiki argues the market misunderstands why Zcash has rallied. He says the real driver is the Tachyon upgrade, which he says developer ZK Dragon targeted for January 2027 (referenced in the episode as "January" / "late January") to ship a shielded pool and post-quantum transparent op codes, scaling toward 50,000 TPS and addressing quantum risk — something he says Bitcoin's roadmap lacks.</p>
<p>He points to Zcash's price bottoming around when Tachyon's blog post was published and argues "smart money" has accumulated around the project, not the privacy narrative most holders cite. He notes that fund managers he's spoken to own Zcash for "privacy" reasons but don't know what Tachyon is, which he treats as a knowledge gap he's trying to exploit.</p>
<blockquote>"I think the real reason Zcash is pumping is because Tachyon... the market doesn't understand Tachyon and that I think is a source of alpha." — Taiki Maeda</blockquote>
<p>He compares this to Zcash's prior Ironwood upgrade on July 28th, after which the price rallied several times over, though he's careful to say he isn't claiming Ironwood directly caused the pump. He plans to consider relonging ZEC perps around December or January, reasoning that catalysts are typically priced in roughly two months ahead of the event, citing LayerZero and Kinetic as past examples of coins bottoming months before their respective launches.</p>

<h3>Bitcoin's "crisis of confidence"</h3>
<p>Taiki argues Bitcoin could trade similarly to how Ethereum performed last cycle — failing to make fresh highs and instead ranging, in his words, possibly for "the next couple of years." He bases this on weak Bitcoin-to-gold and Bitcoin-to-S&P ratios and what he calls unresolved quantum and scalability issues.</p>
<p>He cites growing criticism of Michael Saylor among "OG Bitcoiners" such as Matthew Crader, along with Strategy CFO Andrew King's resignation and the company's removal from MSCI indices, as signs that Bitcoin's narrative momentum is fading. He frames this as part of a broader thesis that marginal capital will rotate from Bitcoin into Zcash as holders "silently" hedge.</p>
<blockquote>"Why do we need Michael Saylor to turn Bitcoin into money with his Wall Street product? Bitcoin is money. We don't need Michael Saylor to turn Bitcoin into money." — a guest view cited by Taiki (Matthew Crader)</blockquote>

<h3>Other signals: PRL, Monero, Zama and the privacy trade</h3>
<p>Taiki highlights PRL as a token he finds interesting but doesn't own, built around the idea that Bitcoin miners pivoting to AI inference weakens Bitcoin's network security while PRL can be mined using the same GPUs. He says it ran from roughly 30-40 cents to near $3 before pulling back, and admits he doesn't fully understand the demand case given AI labs could mine and dump it.</p>
<p>He also points to Monero and Zama as other privacy-related tokens showing strong charts — Monero's chart he calls "insane," and Zama he describes as in "vertical accumulation" — though he stresses he owns neither and flags this could simply be noise following Zcash's own rally rather than independent signal.</p>

<h3>Why he was "punished" for ignoring top signals</h3>
<p>Taiki recounts hosting a Coinbase trading competition in Singapore, where a fan who had made money following his Zcash calls tracked him down with a printed photo of an edited "Chad" version of his face and shouted a $50,000 Zcash price target on camera. He says this happened on October 8th, when Zcash was already down 10%, and jokes that he should have read it as a top signal.</p>
<p>He also flags that stablecoin market cap has been declining, which he calls "a little bit concerning" since it implies dollars leaving the ecosystem, something he says was part of his short thesis on Ethereum and altcoins last year. He notes stablecoin market cap had been "pretty resilient" since October 10th despite this, though recent events like a Binance hack and AI-related exploits may be prompting capitulation.</p>

<h3>Where he goes from here</h3>
<p>Taiki says he's going on a "dopamine detox" after the liquidation, as he's done following prior large wins and losses, before considering relonging Zcash into the Tachyon catalyst and preparing for the Variational airdrop. He maintains that most of crypto, including memecoins and influencer-driven tokens, "should not be touched" and is "worthless," arguing that outsized returns will come from concentrated bets like Zcash and perp-DEX tokens such as Hyperliquid, plus farming new perp exchanges.</p>
<p>He frames overall market sentiment as deservedly poor given grifting and rug pulls, but says he intends to keep focusing narrowly on Zcash and perp-DEX farming rather than chasing broader altcoin moves, acknowledging his reputation as a sometimes "counter signal" while standing by his directional calls over the past six months.</p>

Taiki Maeda gets liquidated on ZEC perps, says he'll relong into January's Tachyon upgrade

After closing a Zcash leveraged long worth millions on a forced sale, Taiki Maeda lays out why he still expects ZEC toward $2,000-plus and argues Bitcoin faces a "crisis of confidence."

Targets above $2,000 next year on Tachyon upgrade, plans to relong into January

Zcash

Could range for years amid quantum and scalability doubts, though he still holds some

Bitcoin

Expects it could flip BNB over the next two years, holds it in portfolio

Hyperliquid

Interesting AI-mining narrative but he doesn't own it and questions the demand logic

Pearl

Chart looks strong but he does not own it

Monero

Expects it to keep ranging low for years, cites prior short thesis

Ethereum

Ziquidated after putting my Net Worth into Zcash.

<h3>Key takeaways</h3>
<ul>
<li>Haseeb Qureshi says AI investing today resembles crypto in 2020, maybe 2021, with "no DD" checks and term sheets written on napkins.</li>
<li>He calls Venice, a privacy-preserving AI inference platform, the asset he has "the most belief in right now."</li>
<li>He frames Zcash as "an insurance policy on your Bitcoin" against both seizure risk and a potential cryptography break.</li>
<li>Dragonfly passed on Hyperliquid's seed round at a $600 million valuation; Jeff later declined offers above $1 billion.</li>
<li>Qureshi says he would turn bearish on crypto if it betrayed its core values or lost its culture of experimentation.</li>
</ul>

<h3>AI mania today looks like crypto circa 2020-2021, Qureshi says</h3>
<p>Haseeb Qureshi, managing partner at Dragonfly, opens by comparing the current AI investing boom to crypto's earlier frenzy. He describes San Francisco AI fundraising as "super social, no DD thing," with "checks flying out the door" and "term sheets written on napkins." He says valuations are being handed to "neolabs that have just four guys and an idea."</p>
<p>Asked when this reminds him of crypto, he places it at "maybe 2020," adding "it gets crazier before getting sad for a while." He points to the scale of AI data-center buildouts and debt, noting Anthropic's compute commitments of "something like 500 billion," and warns that a dip in demand would make "the math" hard to work for labs carrying that obligation.</p>
<blockquote>"It's human nature. It's just like this is what happens when people are really excited about something and they feel they have to get involved." — Haseeb Qureshi</blockquote>

<h3>Winning deals is about access and brand, not just picking right</h3>
<p>Qureshi lays out what he calls a "ladder of skills" in venture: deal flow, analysis, picking, and finally winning. He argues picking is "only halfway up the ladder" because many firms can identify a winner like OpenAI but still lack "a seat at the table." Winning the deal, he says, is the scarce skill.</p>
<p>He explains that investors at large firms such as a16z or Sequoia are "renting the brand" of the firm, while smaller or independent investors must build their own public brand to compete. He calls this "the number one thing that is going to make you successful in VC."</p>
<blockquote>"If you are not at one of those places, then you got to make the brand." — Haseeb Qureshi</blockquote>

<h3>Early bets: NEAR, pivoting from AI founders in 2018</h3>
<p>Qureshi says Dragonfly co-led NEAR's seed round in 2018, when the project was called Near AI and was founded by Ilia, a co-author of the Transformers paper, alongside Alex. He says the founders pivoted from AI to blockchain after concluding Ethereum's engineering was amateurish, and that Dragonfly initially passed on their pitch before the founders returned the next day with a revised protocol addressing every objection.</p>
<p>He says Dragonfly still holds NEAR because "we believe in the team" and the "job's not done," noting NEAR's current focus is as a cross-blockchain exchange allowing trading across chains from one account. He contrasts this with NEAR's original Ethereum-killer sharding thesis, which he says succeeded technically (reaching production before Ethereum's own sharding plans) but didn't drive adoption the way hoped.</p>

<h3>Ethena's seed round came right after Terra's collapse</h3>
<p>Qureshi says Dragonfly led Ethena's seed round in what he believes was 2022, immediately after the Terra/Luna collapse, when "people thought obviously DeFi is not investable now." He says founder Guy had seen a predecessor protocol, Lemma Finance (which Dragonfly also backed), fail because onchain perpetuals lacked enough open interest, and realized that exchange custody practices had shifted enough (via providers like Fireblocks and Anchorage) to let Ethena solve counterparty risk by operating partly offchain.</p>
<p>He says Dragonfly still holds Ethena's token, arguing demand for yield on dollars will grow as the Genius Act bars regulated US stablecoins from paying yield directly, leaving a gap Ethena is positioned to fill "as demand for dollars increases around the world."</p>

<h3>Variational and Hyperliquid: betting on the RFQ model for onchain trading</h3>
<p>Qureshi says Dragonfly backed Variational's seed round on the strength of founder Lucas, even though the firm doesn't usually invest in market makers. After Variational pivoted post-FTX toward an RFQ (request-for-quote) trading model for onchain derivatives, Dragonfly led a later round reportedly at $50 million. He argues RFQ is the only viable model for scaling real-world-asset trading onchain, since "there's no way" to build deep onchain order books for thousands of traditional-finance symbols.</p>
<p>On Hyperliquid, Qureshi says Dragonfly pitched founder Jeff at a $600 million valuation pre-TGE and was turned down, as Jeff declined offers from several VCs, including some above $1 billion later. He says this self-funding conviction is "the number one thing you want to look for in a founder." On taking profits, he says Dragonfly's roughly ten-year fund cycle gives latitude to hold winners, citing Polymarket, where Dragonfly invested at a $175 million valuation while a reported new round values the project near $21 billion.</p>

<h3>Top current conviction: Venice for AI privacy, Zcash as "insurance" on Bitcoin</h3>
<p>Asked for the asset he believes in most right now, Qureshi names Venice, a privacy-preserving AI inference platform growing "10% month over month" and now "well over 100 million revenue run rate," up from roughly 15 million when Dragonfly first spoke to them. He argues AI chat logs currently carry no legal privacy protection in the US under the "third party doctrine," meaning logs can be subpoenaed without a warrant, and frames Venice as building "that VPN for AI."</p>
<p>On Zcash, he describes it as "a privacy preserving version of Bitcoin" and repeats a framing he attributes to Naval that it functions as "an insurance policy on your Bitcoin," both against seizure and against a potential break in public-key cryptography, referencing OpenAI's reported mathematics results and speculation about AI-assisted cryptography research. He suggests sizing a Zcash position in proportion to one's perceived risk in Bitcoin, though he avoids giving a specific allocation.</p>
<blockquote>"The more you worry, the bigger your insurance policy ought to be." — Haseeb Qureshi</blockquote>

<h3>What would make him bearish on crypto</h3>
<p>Qureshi, describing himself as someone who stayed optimistic through recent bear markets, says two things would shift him toward bearishness: a "betrayal of our core values," meaning crypto giving up on being an alternative to existing systems, and a loss of the industry's culture of experimentation. He explicitly rejects the idea that institutional adoption and ETFs (including products wrapping Zcash) dilute cypherpunk values, arguing institutions are "running on the cipherpunk rails," not the reverse.</p>
<p>He contrasts Sequoia's continued crypto investing despite its FTX exposure with other FTX-adjacent investors who left the space entirely, calling that persistence "the essence of Silicon Valley."</p>
<blockquote>"That culture is a very important and sacred culture. And I think if you lose that, that would make you more bearish on crypto." — Haseeb Qureshi</blockquote>

Dragonfly's Haseeb Qureshi: AI funding mania mirrors crypto's 2020, names Venice and Zcash as top picks

The Dragonfly managing partner compares today's AI investing frenzy to early crypto cycles and lays out his theses on Near, Ethena, Variational, Hyperliquid, Venice and Zcash.

Frames it as insurance against Bitcoin seizure risk and cryptography breaks

Still held since 2018 seed; believes it can get much bigger as cross-chain exchange

NEAR Protocol

Expects demand for dollar yield to grow as Genius Act bars yield-bearing stablecoins

Ethena

Dragonfly was rejected as investor; sees total addressable market as the global CFD market

Could face seizure or cryptography risk, which Zcash is meant to hedge

$4 Billion Fund Manager Reveals The 5 Crypto Coins He’s Going All In On In 2027 | E194

<h3>Key takeaways</h3>
<ul>
<li>A Senate Banking Republican, Kevin Kramer, called for Jensen Huang to resign over his relationship with an alleged Chinese spy.</li>
<li>Polymarket odds that Huang leaves before July next year sat at 38%, roughly unchanged from the prior day.</li>
<li>Coreweave shares rose 26.8% after revenue beat guidance, helping lift the Nasdaq and revive the AI trade.</li>
<li>Anthropic's IPO valuation talk moved from a 1.26-1.41 trillion range toward a reported 1.5 trillion being weighed by bankers.</li>
<li>Jeff Yass argued stock market bubbles do not exist, saying price drops reflect "uncertainty resolving," not delusion correcting.</li>
</ul>

<h3>Jensen Huang's scandal draws resignation calls while the board holds steady</h3>
<p>The hosts discussed continuing fallout from reports that Nvidia CEO Jensen Huang had a relationship with Alice Wong, described in reporting as a suspected Chinese spy, with Axios reporting that Commerce Secretary Howard Lutnick was sent a classified memo warning about the relationship. The hosts noted this raises questions about how the department weighed security risks when clearing Nvidia chip sales to China-affiliated entities. North Dakota Senator Kevin Kramer, a senior Republican on the Senate Banking Committee, called for Huang to step down, saying the potential for extortion given Nvidia's position is "so tremendous" that "I just think he's got to go."</p>
<p>The hosts noted Polymarket odds that Huang leaves the CEO role before July of next year sat at 38%, largely unchanged from the previous day, and said he appears to be in the "business as usual, not leaving" camp. CNBC reported Huang addressed the matter briefly in a Monday morning executive meeting, apologized, took no questions, and moved on to business topics. The hosts speculated about succession technicalities, comparing it to Elon Musk's multiple titles at Tesla and SpaceX, and joked a resolution could see Huang become a "techno king" of Nvidia rather than fully departing.</p>

<h3>Debate over Nvidia's CUDA moat intensifies alongside the scandal</h3>
<p>The hosts tied the scandal to separate commentary that Anthropic views all compute as "fungible," noting that Mythos reportedly took three days to get itself running on AMD hardware versus a prior six-to-12-month timeline. One host called this "a 30x, 50x, 100x increase in speed," while noting debate remains over whether Anthropic is actually seeing a 2x increase in AI development. The hosts argued Anthropic's worldview is fundamentally at odds with Nvidia's approach to exporting technology, with one host saying "you had to be bearish on CUDA" once you understood the implications of fast software generation, comparing Nvidia's position to Google helping create large language models that could ultimately disrupt its own business.</p>

<h3>AI trade rallies as Coreweave beats and Anthropic's IPO valuation climbs</h3>
<p>The hosts described the S&P 500 surging above 7,600, roughly 3% from a record high, with the Nasdaq erasing the prior week's losses on Coreweave results and rising expectations for Anthropic's IPO valuation. Nvidia itself was up roughly 4% on the day discussed but remained down 22% over the trailing month. Coreweave shares jumped 26.8% after revenue growth exceeded guidance, though the hosts noted the stock would still need to roughly double to reach its earlier peak. On Anthropic, the hosts cited a Wall Street Journal report of a targeted valuation range of 1.26 to 1.41 trillion, "far below two trillion," alongside an FT report that bankers were weighing raising it to 1.5 trillion. The hosts also flagged a divergence between Ramp's enterprise data showing roughly 1% growth and Anthropic's own claim of around 13% growth, and noted the GDPX benchmark from semi analysis moved from about 12% to 16%, still "a ways to go" before saturation.</p>
<p>The hosts also discussed Jeff Yass, described as founder of Susquehanna, arguing there is no such thing as a stock market bubble, writing that price falls reflect "uncertainty resolving" rather than delusion correcting. <blockquote>"Why would rational investors suddenly lose their mind?"</blockquote> — Jeff Yass, as quoted by the host. The hosts did not take a clear side on Yass's argument themselves beyond noting it "warmed their heart."</p>

<h3>Hosts debate whether top founders have hobbies</h3>
<p>The hosts discussed a claim attributed to Masha Butcher that "the best founders I backed had no hobbies, only one obsession, the company," contrasted with examples like Mark Zuckerberg's jiu-jitsu, Jeff Bezos's sailing and ocean exploration, and Steve Jobs's meditation and long walks. One host argued these examples mostly reflect hobbies picked up after companies scaled and founders gained time and flexibility, rather than during early, pre-product-market-fit grind. <blockquote>"I think I'm getting more convinced that the butcher might be right."</blockquote> — a host. The hosts proposed a working definition that a true hobby must be non-monetizable, involve constant improvement, and require visiting a physical specialty store where "the guy knows your name," disqualifying things like casual gym visits or movie-watching as hobbies without that deeper commitment.</p>

Nvidia board faces pressure as Jensen Huang spy scandal deepens, but AI trade rallies anyway

Hosts cover calls for Jensen Huang's resignation, a resurgent AI trade on Coreweave and Anthropic news, and a debate on whether top founders have hobbies.

Jensen’s Scandal Tests Nvidia’s Board, AI Trade Roars Back, The Great Hobby Debate | Diet TBPN

<h3>Key takeaways</h3>
<ul>
<li>James says the AI selloff was the biggest dollar-value crash since COVID March 2020, followed by a fast V-shaped recovery.</li>
<li>He flags Micron up 12.29%, AMD up over 21% in two days, Marvel up 11%, and Broadcom up 6% in the rebound.</li>
<li>Bitcoin and Solana pierced their 200-day moving averages; Bitcoin only rose 0.61% versus AI stocks' much larger bounces.</li>
<li>James says Tom Lee "lost $6 billion on his ETH trade" and still expects ETH to reclaim 2730, a 35% gain from $2,000.</li>
<li>He calls Micron, Marvel, copper and SpaceX buys on mean-reversion signals, and flags MicroStrategy's MSCI decision as a near-term risk.</li>
</ul>

<h3>AI stocks crash then V-shape back while crypto lags</h3>
<p>James opens by describing the past few days as "bonkers," noting a sharp AI selloff that he calls the biggest dollar-value crash since COVID in March 2020, followed by an unusually fast rebound. He credits CoreWeave's earnings and a "135 billion" backlog with lighting the fuse, alongside memory stocks like Micron rising 12-15%. He lists the day's "monster moves": Micron up 12.29%, Marvel up 11%, Alab up 9%, AMD up 7% (over 21% across two days), Broadcom up 6%, with TSM, Tesla and Nvidia up nearly 4%. By contrast, he notes Bitcoin only rose 0.61% and ETH 1.4%, while SpaceX was roughly flat.</p>
<p>He also grades his prior week's calls, saying he "missed" on Bitcoin and Solana, which he didn't expect to fall as hard as they did, and says Nvidia got hit by what he calls a "spy scandal" that technical analysis couldn't have flagged: He says his calls on AMD, Broadcom, Marvel, Lumentum, Light, Alab, SpaceX, Micron, Google and Palantir were correct.</p>

<h3>Macro backdrop: rates, VIX and oil</h3>
<p>James walks through the US 10-year yield, which fell from near 5.4% to 4.8% before rebounding, which he calls "bad" given some Fed officials reportedly want to raise rates twice before year-end, which he calls "ridiculous" given oil's influence. He expects the VIX to mean-revert lower after spiking above 28-29, noting it was at 17.48 with support near 14.46 and a 200-day moving average around 18.</p>
<p>On oil (WTI), he points to a "clear double top" near $106 and says price is back down toward $85-86, adding it could rise again "unless they can get peace in Iran," which he is skeptical of given repeated false signals. He also discusses the Nasdaq piercing its 200-day moving average similarly to April 2026, with a "magnificent" V-shaped rebound, while the S&P 500 did not hit its 200-day line, which he attributes to lighter tech exposure.</p>

<h3>Crypto: Bitcoin, Ethereum and Solana still below key levels</h3>
<p>James says Bitcoin pierced his "level five" at 77K and looks ready to turn back up from the 73-74K area, arguing it is "too cheap down here" given money printing and deficits. He notes Bitcoin versus gold "got hammered." On Solana, trading near 83 and deep under its 200-day moving average (around 89-90), he says it "looks ready to turn back up," adding that if crypto has momentum at all, Solana "should be a lot higher," citing a figure of "200 bucks, 300 bucks," though he ties this to AI/crypto convergence that hasn't happened yet.</p>
<p>On Ethereum, he references Tom Lee's ETH trade losing "$6 billion" and dismisses bullish ETH targets: <blockquote>"A lot of people who tell me said ETH is going to $25,000... not going to happen."</blockquote> He flags a buy signal under ETH's 200-day moving average and says a move back to resistance near 2730 is plausible, a 35% gain from the current $2,000 level. He also flags MicroStrategy's pending MSCI decision at "11:00 p.m. Eastern," saying removal could pressure the stock short-term but that the asset's value mainly tracks its Bitcoin holdings, which he puts at "over 850,000 Bitcoin."</p>

<h3>AI and commodity names: buy signals and caveats</h3>
<p>James runs through individual tickers using his mean-reversion model: he calls SpaceX's dip to 120 a buy signal with a "70-75%" historical win rate, Tesla a "no-brainer" buy below $350 or at the $335 level it hit, and Nvidia a buy around 170-174 despite the unexplained "spy" drop, saying "the fundamentals are not changing" given CUDA and sold-out Vera Rubin chips. He flags Micron's dip as the best since April 2026, targeting "$1,000," then "1,200," then "1,500-1,600 within a year or less," linked partly to Tesla's Optimus robot memory needs.</p>
<p>He also calls Marvel a "clear buy" near 120-130 following its Lumentum acquisition-related selloff, says Broadcom's "330 kill zone" still holds, and flags Palantir's mean reversion as the worst since March 2026 before heading back toward "200 bucks." On commodities, he says gold shows a buy signal near 3920-3960, silver shows no buy signal at its cheapest level in "13 months," and copper looks oversold with a buy around "$6.19" targeting "$6.90." He stresses leverage discipline throughout, saying he never uses more than 20% of his margin.</p>

InvestAnswers: AI stocks V-shape back as crypto "gives it all back"

James walks through 21 charts showing AI names snapping back hard from a crash while Bitcoin, Ethereum and Solana stay under their 200-day lines.

Sees it too cheap near 73-74K, expects a turnaround given money printing and deficits

Sees a technical buy signal toward 2730 but dismisses $25,000 bullish targets as not happening

Expects a turn back above 200-day average near $89-90, could reach 200-300 if AI/crypto converge

Solana

21 Charts: AI Up, Crypto Down + What Smart Money Knows Right Now📈📉

<h3>Key takeaways</h3>
<ul>
<li>Tyler Shultz says Theranos had only seven validated tests at his departure, far short of Elizabeth Holmes's claimed 300.</li>
<li>He describes a syphilis test with roughly 65% then 80% sensitivity and over 20% false positives among colleagues, yet it was approved anyway.</li>
<li>Shultz says lawyers led by David Boies ambushed him at his grandfather George Shultz's house with a restraining order and a perjury-risk affidavit.</li>
<li>He spent an estimated $400,000 to $500,000 on legal fees and says he was followed by private investigators even a year after leaving.</li>
<li>His grandfather, a Theranos board member, never fully believed him and the board backed Holmes "100%" even after Wall Street Journal exposés.</li>
</ul>

<h3>A Stanford junior falls under Elizabeth Holmes's spell</h3>
<p>Tyler Shultz first met Elizabeth Holmes at his grandfather George Shultz's house while he was a junior at Stanford. He describes her as hypnotic, saying she wore a black turtleneck modeled on Steve Jobs and had "piercing blue eyes" and a voice that made a room go silent. Holmes pitched a vision in which a single drop of blood could run any test a central lab could run, usable in battlefields, helicopters and patients' homes. Shultz says he was "completely sold" and switched his college major from mechanical engineering to biology at her urging.</p>
<p>During an internship, he noticed colleagues who had worked there for years had never seen the actual product, and the lab operated under strict silos, locked doors and NDAs justified by fears of corporate espionage. A colleague warned him on his last day as an intern, "You shouldn't come back here. You should go somewhere else and learn how to do real science."</p>

<h3>Inside the lab: a product that wasn't being used</h3>
<p>Shultz's first day as a full-time employee coincided with Theranos's public product launch at Walgreens. He says that despite the launch, zero tests had been validated on the actual Theranos device; patient blood was drawn by needle, not finger prick, and tested on third-party machines from Siemens or other vendors. He recalls a validation-team manager quitting and being found crying in the parking lot on his very first day.</p>
<p>Employees worked around the clock in a lab nicknamed "Normandy," with Sunny Balwani conducting 3 a.m. walkthroughs to ensure people were working. Shultz says quality-control failures were routinely overridden: He also says Theranos locked the door to the room with its real devices during a regulatory inspection so inspectors only saw the room with third-party equipment, nicknamed "Jurassic Park."</p>

<h3>The syphilis test and a culture built on fear</h3>
<p>Shultz says a pivotal moment came during syphilis test validation: known-positive samples returned about 65% sensitivity, which was deleted and retested to get 80%; testing presumed-negative samples, including employees themselves, produced over 20% false positives. Despite this, a doctor signed off and the test moved to real patients. He says the culture of fear had been built deliberately over roughly a decade, with disagreement met by firing, and firings sometimes followed by lawsuits publicized company-wide as a warning. He also says employees received no stock options, which, combined with mortgages, visas and family obligations, discouraged people from quitting despite widespread private acknowledgment that the technology didn't work.</p>

<h3>Raising concerns and a calculated resignation</h3>
<p>Shultz emailed Elizabeth Holmes directly with data-backed concerns, trying to frame himself as a team member rather than an accuser. Holmes forwarded it to Sunny Balwani, whose reply, in all red and partly capitalized, called Shultz "arrogant, ignorant, patronizing, reckless" and said he had no understanding of basic math or statistics. Shultz resigned with two weeks' notice. On his last day he secretly printed around 200–300 copies of his email exchange, smuggling a ream of paper out under his sweatshirt, and walked out unescorted because security was "busy" — a detail that later contradicted Theranos's claim that he had left in an angry, dramatic confrontation.</p>
<p>He says his grandfather repeated Theranos talking points almost verbatim, including an unverified claim attributed to Jim Mattis that the device was used on medevac helicopters in Afghanistan with a "98% survival rate, up 30% from 68%." Shultz does not know whether Holmes fabricated the quote but says his grandfather relayed the same specific claims to him directly.</p>

<h3>The ambush: lawyers, a fake NDA and a pencil-written promise</h3>
<p>After a Wall Street Journal reporter contacted him, Shultz agreed to speak on deep background. When Theranos learned of the investigation, his grandfather arranged to meet him alone, but two Boies Schiller lawyers were already waiting in the house. Instead of a one-page NDA as promised, Shultz was presented with a temporary restraining order, a notice to appear in court, and an affidavit requiring him to deny being a source and to name other sources under penalty of perjury. He refused, and after prolonged argument his grandfather wrote "Theranos will not sue Tyler forever" in pencil at the bottom of the document — which the lawyers accepted, apparently confident the pledge, in pencil, carried no real weight, or that Shultz would be forced to recant soon after.</p>
<p>Shultz delayed signing by requesting independent counsel, eventually reaching his grandfather's estate attorney, who recognized the opposing lawyer's name and reacted with alarm, telling him David Boies was known as "the number one gun in corporate America." Shultz says he ultimately spent $400,000 to $500,000 on a full legal team and says the negotiation cycle was repeatedly drawn out, which he came to see as a deliberate war of attrition.</p>

<h3>Isolation, fear and a slow recovery</h3>
<p>During the months before publication, Shultz says he cut off contact with friends, blocked every number in his phone, and communicated mainly over Snapchat because the messages disappeared. He describes sleeping with a knife near his bed, fearing Theranos or Balwani could physically target him, and says he was told private investigators watched him roughly 80 to 90% of the time he was in public. He says a lawyer hired by his parents had a briefcase of case documents stolen from her car during this period, which he does not believe was coincidental. He calls this period the lowest point of his life and says he was likely clinically depressed, though he did not recognize it then.</p>
<p>When the Wall Street Journal's first article ran, Shultz was initially disappointed, feeling it had "pulled a lot of punches," but it proved to be the start of a sustained series of reports. Subsequent regulatory inspections by CMS and the FDA found practices that put patient health "in immediate jeopardy," and Holmes was barred from operating a lab for two years. Roughly a year later, with his grandfather still on the Theranos board and the board publicly backing Holmes "100%," Shultz decided to go on the record with the Journal, against his legal team's and parents' advice, saying he wanted to force accountability rather than continue hiding.</p>

<h3>Aftermath, reflections on gender and advice for founders</h3>
<p>Shultz and the hosts discuss how Holmes leveraged her identity as a prominent female founder, with Shultz arguing, "She could not achieve what she was able to achieve had she been a man," while also saying she later used being a new mother as a sympathetic narrative during sentencing. He says his personal life improved steadily after going public, including reconnecting with a close friend who became his wife. He urges employees who witness wrongdoing to seek legal counsel early, noting that government whistleblower programs, such as the SEC's, can provide free legal representation on contingency. He advises founders to define core values explicitly and treat small ethical lapses seriously, arguing this builds a culture that surfaces problems "before they really grow into these huge issues."</p>

Tyler Shultz details Theranos cover-ups, gaslighting and a David Boies legal ambush

Theranos whistleblower Tyler Shultz recounts faked tests, a family betrayal and the legal pressure campaign that followed his Wall Street Journal tip, offering lessons for founders and employees on culture and speaking up.

Inside Theranos with ex-Employee Tyler Shultz

<h3>Key takeaways</h3>
<ul>
<li>KoolKrypto says his top-100 Hyperliquid ranking comes mostly from avoiding blowups, never using more than 1x leverage.</li>
<li>He sees Zcash as "ungated by revenue," able to trade like a memecoin to many multiples, but warns it lacks a fundamental price floor if sentiment turns.</li>
<li>He expects lighter to outperform Hyperliquid over the next 3-6 months on regulatory grounds, but trusts Hyperliquid's team more long term.</li>
<li>Derive dominates options volume now, but he thinks its market share could fall to 55-60% even as the whole options pie grows 5-10x.</li>
<li>He argues crypto is at "maximum risk, maximum reward" as AI-bubble fears and a cooling rate environment could redirect risk-chasing capital into crypto.</li>
</ul>

<h3>Avoiding blowups, not chasing home runs</h3>
<p>KoolKrypto, described in the episode as a top-100 Hyperliquid perp P&L trader, tells David Hoffman that his ranking is less about catching huge trades and more about risk control. He says he "literally never" uses more than 1x leverage across his portfolio and that most top traders on the leaderboard simply avoided large losses rather than capitalized on huge wins. He notes some of his perp P&L actually came from hedging Hyperliquid spot holdings with local-top shorts rather than any single giant directional bet.</p>
<p>Asked whether leaderboard data is useful signal, he says it has "a little bit of signal" but is often distorted by delta-neutral funding-rate arbitrage funds being mistaken for directional whales. He says he spends more time watching large position holders after he already has exposure to a ticker, rather than using leaderboard-watching to find new trades.</p>

<h3>The "revenue meta" and how he values tokens</h3>
<p>KoolKrypto says his background is DeFi-native, having started a fund years ago betting on a rotation into revenue-generating tokens — a thesis he admits he was early and wrong about for a long time before 2026 vindicated it. He now argues midterm election results could spark a rotation out of revenue-meta tokens, which he says are more exposed to regulation, and into "store of value" tokens like Bitcoin and Zcash, which benefitted from a broader move he compares to a gold rally.</p>
<p>He pushes back on a strict binary between "money" tokens and "revenue" tokens, citing Ethereum as a token that doesn't fit cleanly into either bucket.</p>

<h3>The bull and bear case for Zcash</h3>
<p>KoolKrypto argues Zcash currently sits in "the perfect setup" narrative-wise, combining privacy and quantum-resistance themes, but says it is "such a crowded consensus crypto trade" that it needs genuine external capital, not more crypto insiders, to re-rate higher. He points to wealthy individuals outside crypto, citing Naval as a figure associated with Zcash interest, as the kind of buyer that could move the asset without needing mass retail adoption.</p>
<p>Hoffman counters that Zcash could be framed as a percentage of Bitcoin's market cap, arguing Bitcoiners occasionally defect into new assets, as some did into Ether in 2021, and that it wouldn't take much capital to push Zcash to "three, four, five%" of Bitcoin's value. KoolKrypto agrees but flags that it's hard to know how much of that defection has already happened.</p>
<p>He also raises a bear case: Zcash's code isn't as "ossified" as Bitcoin's, raising attack-surface risk, as seen in what he calls "the whole Orchard fiasco," and that it lacks a fundamental price floor the way revenue-generating, buyback-funded tokens do. <blockquote>"Zcash can't really just kind of go up at like a 10% kagger... it's kind of just either going to like step function up... or it's just kind of going to bleed out."</blockquote> — KoolKrypto. He says he expresses his Zcash view through options rather than perps, because options let him target large multiples ("10,000," "5,000" are cited as levels others discuss) without liquidation risk during the expected 60-70% drawdowns along the way.</p>

<h3>Valuing exchange tokens: buybacks change the math</h3>
<p>KoolKrypto says he finds the most edge in revenue-generating "decks" tokens — Hyperliquid, Lighter, Derive — because many investors mis-contextualize why they trade at different multiples. He argues Hyperliquid's HIP-3/USDC-linked revenue is more durable, like treasury yield, than Lighter's, which he says is one reason Lighter trades at a lower multiple. He also says Pump's revenue is discounted because markets assume meme-coin revenue cannibalizes its own future.</p>
<p>He says tokens with aggressive buybacks relative to revenue, such as Hyperliquid, Derive, Lighter and Arrow, don't need broad recognition to re-rate because cheap buybacks automatically manifest in price: "hype is buying three times as much hype back when it was trading at you know 30 bucks than when it's trading at 90 bucks." He adds that fundamental data rewards being early and contrarian, unlike technical levels, where he says it pays to watch the same signals everyone else watches.</p>

<h3>Hyperliquid vs Lighter</h3>
<p>KoolKrypto calls Lighter "a better hold for the next 3 to 6 months," partly because he views Hyperliquid as having been somewhat politically aligned with the Trump administration, which he suggests could shift under changing political winds, while Lighter may have a cleaner path to regulatory clarity. He flags a possible bear case for Lighter too: the untapped US retail perp market may be smaller than assumed, since users already access offshore venues via VPN.</p>
<p>Longer term, he says he trusts the Hyperliquid team more and would rather hold hype in a multi-year coma trade than lighter's token, comparing Hyperliquid's potential trajectory to Binance's offshore dominance. He stresses that liquidity depth is a bigger moat than feature superiority, noting he still trades mostly on Hyperliquid despite preferring Lighter's UI.</p>

<h3>Options as the next growth frontier</h3>
<p>KoolKrypto says perp markets are closer to saturation, citing the lingering hit to open interest and volume from "1010" (an event he says killed a lot of perp OI), while options adoption is on "a straight hockey stick" trajectory. He argues even derive bulls underestimate how much the options market could grow, suggesting it "could very easily 5 to 10x next year," though he doubts Derive keeps its current roughly 90-95% market share of options notional as competition from Lighter, Paradex's Dime, Hyperliquid's Synapse, and Variational Pro intensifies.</p>
<p>He says Hyperliquid's architecture isn't well suited to native options and expects the team to eventually enter the market, but likely 9-12 months behind Lighter, which he says is explicitly building options alongside its Robinhood partnership.</p>

<h3>Macro setup: "maximum risk, maximum reward"</h3>
<p>Looking ahead to the rest of 2026, KoolKrypto says crypto sits at a "peak career risk" moment for institutional allocators: a potential cooling of the AI bubble could free up risk-seeking capital that previously favored semiconductors and AI stocks over crypto. He notes Bitcoin is still trading well below its all-time high, unlike broader tech, implying crypto already front-ran part of its own correction.</p>
<p>He frames the key open question as whether the revenue meta has topped in favor of store-of-value trades like Zcash, or whether both Zcash and hype are simply the most crowded, vulnerable assets if crypto falls further. <blockquote>"If you're right, you're right gigantically and if you're wrong, you're also kind of wrong gigantically."</blockquote> — KoolKrypto, leaving the direction of the next three to six months explicitly unresolved.</p>

Hyperliquid top trader says options market could grow 5-10x crypto-wide next year

KoolKrypto, a top-100 Hyperliquid P&L trader, lays out his risk-first approach, the Zcash bull and bear case, and why he sees options eclipsing perps as crypto's next growth market.

Trusts the team long term, expects multiples higher eventually but not a 10x from here

Debated as the ultimate store of value; Bitcoin could rally most after a crisis that also hits Zcash

Currently dominant options venue but expects market share to fall to 55-60% as options volume grows 5-10x

Derive

Mentioned as an example of a token some traders have become associated with huge singular perp trades

Trades cheap on revenue multiples because market assumes current revenue cannibalizes future revenue

Pump.fun

How a Top 100 Hyperliquid Trader Finds the Next Big Trade

<h3>Key takeaways</h3>
<ul>
<li>The hosts detail seven new allegations about Nvidia, Jensen Huang, and Alice Wong, including alleged interference in a federal probe.</li>
<li>Prediction markets put the odds of Jensen Huang exiting before July at almost 40%, up from around 30-31% the day before.</li>
<li>The hosts call the scandal possibly "the largest unforced error in business history," eclipsing even Microsoft's market cap.</li>
<li>Anthropic is reportedly on track for an IPO next Thursday, with new filings showing an $80 billion ARR figure cited by a guest.</li>
<li>Amazon's Ring unveiled a $250 flying security drone and a $1,000 roaming robot, which the hosts debate as either promising or risky given recent inference-provider hacks.</li>
</ul>

<h3>Alice Wong allegations multiply, with Jensen Huang's position in question</h3>
<p>The hosts walk through seven new allegations surrounding Nvidia, CEO Jensen Huang, and Alice Wong, cautioning that reporting comes from multiple outlets including Wired, Bloomberg, and The Information and should be treated cautiously. Claims include that Nvidia's general counsel contacted the Commerce Department about an investigation into Megaspeed smuggling chips into China, that Nvidia's counsel called a Commerce undersecretary after Alice Wong was detained in Singapore, and that Nvidia undercounted roughly 50,000 GPUs sold to Megaspeed. The hosts also note the FBI, Congress, and Nvidia's own board are now investigating, though it is unclear whether this leads to public congressional testimony.</p>
<p>The hosts speculate Jensen Huang could become a political target given his visibility as an AI-administration ally. One host says, <blockquote>"They're going to throw the book at him. It's going to be the Girardian scapegoat."</blockquote> They compare the episode to historical business scandals, concluding it may be unprecedented in combining personal scandal with geopolitical stakes, "eviscerating hundreds of billions of dollars of market cap."</p>
<p>Tyler reports that prediction markets now price Jensen Huang's departure before July at <strong>almost 40%</strong>, up from around 30-31% the day before. The hosts say there's no clear succession plan, calling it "just a crisis."</p>

<h3>The corporate web behind Alice Wong and Megaspeed</h3>
<p>The hosts reference a detailed breakdown from The Information tracing a complex ownership structure linking Alice Wong, her husband Kong Wei, and companies including Shanghai Century Leong, Shanghai Century Longtang, and Seven Road Holdings, which formerly owned Megaspeed. They note Shanghai Century Leong was founded in June 2023 and by October had become a preferred Nvidia cloud partner, facilitating over <strong>$7.5 billion</strong> in Nvidia chip shipments, including banned chips.</p>
<p>One host raises the possibility Nvidia was deceived rather than complicit, framing a best-case defense: Nvidia's teams checked facilities and believed chips would stay local. But he notes <blockquote>"there's been a year since that detention where they could have been front-running this story,"</blockquote> undercutting that defense. A colleague described Alice Wong as telling customers she could "always get the supply, citing her close relationship inside Nvidia and with distributors."</p>

<h3>What a Jensen exit could mean for the AI buildout</h3>
<p>A guest argues much of Nvidia's moat stemmed from Jensen Huang's personal dealmaking and willingness to backstop neoclouds and labs, warning that his legal troubles could force a pause in that activity. He says VCs have treated Nvidia as a backstop: <blockquote>"shoot for the moon but even if you miss, Jensen will give you another up round."</blockquote> The hosts discuss Bill Bishop's claim that Nvidia executives ran a "smear campaign" against export-control proponents, with Bishop predicting more revelations ahead. Travis Kalanick posted guidance framed as general advice on surviving a "coup," widely read as referencing Huang.</p>

<h3>Anthropic's IPO track and Fable 5.2 release</h3>
<p>The hosts note Anthropic has filed its prospectus ahead of an IPO targeted for next Thursday, timing that coincides with the Nvidia scandal. A guest cites an <strong>$80 billion</strong> ARR figure and argues the scandal created a "golden opportunity" for investors given a cut valuation. The hosts highlight that Anthropic's site offers retail investors direct brokerage links (Schwab, Robinhood, E-Trade) to participate in the IPO, which one host calls unusually "down to earth" relative to typical large IPOs reserved for big banks.</p>

<h3>Ring's new security drone and inference-provider hacks</h3>
<p>Amazon's Ring Robotics unveiled a $250 flying security camera for homes and a $1,000 roaming robot for businesses, with the flying camera mapping rooms and investigating triggered alarms. The hosts debate indoor versus outdoor use cases and raise cybersecurity concerns given Amazon's scale, with one host noting Amazon "can spend more on cybersecurity for this product than any other probably security company."</p>
<p>Separately, the hosts discuss a hack at inference provider Fireworks, noting attackers were not able to exfiltrate all model weights and the breach was caught the same day. They reference a September comment from Ilya Sutskever warning that neoclouds have limited cybersecurity, which a host says "wasn't quite" validated here since full compute takeover didn't occur, though data and weights were stolen.</p>

Market Feed