• Treat NVIDIA (NVDA) as a governance and regulatory risk: avoid assuming the allegations are short-lived, and monitor investigations, export-control changes, and leadership developments before making a move.
• Consider Anthropic’s reported IPO only after reviewing its offering valuation and fundamentals; the cited $80 billion ARR figure and valuation cut are not enough to establish an attractive entry price.
• Be cautious with CoreWeave (CRWV): NVIDIA’s past support is no guarantee of future funding, and financing costs may constrain its infrastructure expansion.
• Track tighter U.S. controls on advanced chips and cloud access as a potential sector-wide risk, especially for companies serving China-linked customers.

NVIDIA (NVDA)

• The discussion was strongly cautious about NVIDIA, describing the allegations surrounding CEO Jensen Huang and Alice Wong as a serious, potentially prolonged overhang on the company.
  • The hosts discussed reports of possible interference with a federal investigation, intervention after Wong’s detention, and roughly 50,000 additional GPUs allegedly sold to Megaspeed. They stressed that the claims vary in reliability and remain under investigation.
  • NVIDIA’s board, the FBI, Congress, and other authorities were said to be examining aspects of the situation. The hosts expected further scrutiny and suggested the story could take months or years to resolve.
  • They raised the possibility of Huang leaving his role, while noting that NVIDIA lacks an obvious succession plan.
• The episode highlighted possible consequences for NVIDIA’s business and strategy: tougher export controls, a reset in relations with Congress, and reduced ability to fund or support AI companies and infrastructure projects.
  • Chris Miller noted that chips allegedly smuggled into China or supplied through Southeast Asian data centers represent only a few percentage points of NVIDIA revenue by the best estimates discussed. The issue could still have outsized regulatory and reputational consequences.
  • The hosts also questioned whether uncertainty around Huang’s leadership could weaken confidence in NVIDIA’s deal-making and in the AI companies that rely on its support.

Takeaways

• Treat the situation as a material governance and policy risk, not just a short-lived headline. Monitor investigations, regulatory responses, and any leadership developments.
• The discussion offered no valuation or price target. Any investment assessment should account for the possibility that China-related sales and NVIDIA’s broader relationships with AI companies could be affected, even if the direct revenue exposure is limited.

Anthropic (Private; IPO discussed)

• The hosts said Anthropic was reportedly preparing for an IPO the following week and cited an $80 billion ARR figure.
  • They said the company had cut its valuation, which they viewed as a potentially more attractive entry point for investors.
  • They described Anthropic’s business as continuing to work despite the market and geopolitical turmoil, and expressed optimism about investor access to the IPO.
• The hosts also noted that NVIDIA-related news was overshadowing Anthropic’s IPO coverage.

Takeaways

• The discussion presents Anthropic as a potentially notable AI investment opportunity, but provides no IPO price, valuation details, or financial breakdown beyond the ARR figure.
• Before assessing the offering, compare the IPO valuation with the company’s growth and business fundamentals; the transcript’s optimism is not a substitute for that analysis.

Microsoft (MSFT)

• The hosts noted that Microsoft had a greater market capitalization than NVIDIA at the time, describing the change as striking given NVIDIA’s recent position as the world’s largest company.

Takeaways

• The comparison is useful market context, but the episode did not offer a specific view on Microsoft’s valuation or prospects. It should not be read as a recommendation to buy or sell either company.

CoreWeave (CRWV)

• One host said CoreWeave had been “beat up” and recalled previously believing Jensen Huang would not let the company fail.
  • The host said the current uncertainty around Huang and NVIDIA had weakened that assumption.
  • The discussion also described capital and financing costs as constraints on the AI infrastructure buildout, including data centers and related facilities.

Takeaways

• Avoid treating NVIDIA’s past support as a guarantee of CoreWeave’s future funding or success. Consider the company’s ability to finance its infrastructure and sustain demand independently of any one partner.

AI Infrastructure and Semiconductor Policy

• The guests discussed a possible shift toward tighter U.S. controls on advanced chip sales and cloud access, following the allegations involving NVIDIA and China-linked companies.
  • Chris Miller said NVIDIA may need to refocus on the U.S. market and improve its relationship with Congress.
  • He also flagged the potential impact on data centers outside China that serve Chinese customers, including Southeast Asian providers and Oracle’s relationship with ByteDance.
  • Google, Amazon, and Intel were mentioned as examples of companies with different approaches to China, but the episode did not provide investment recommendations on them.

Takeaways

• Treat export controls and scrutiny of cloud services as a sector-wide policy theme, not solely an NVIDIA issue. Companies exposed to advanced-chip sales or data-center services for Chinese customers may face changing rules.
• The episode did not identify specific winners or provide price targets; track policy developments before drawing conclusions about individual companies.

EBS (Private)

• EBS, a defense-technology startup, announced a $30 million seed round and said it had generated seven-figure revenue within its first six months.
  • The company uses off-the-shelf robots and AI to help automate defense-factory tasks, beginning with wire-harness assembly.
  • Its approach is aimed at lower-volume manufacturing, where flexible robots may be more useful than dedicated production lines. The company said it expects revenue to cover its capital-equipment costs, with funding directed mainly toward AI talent and specialized training data.

Takeaways

• The discussion points to a possible opportunity in defense manufacturing automation, especially for factories producing varied designs in relatively small quantities.
• The company is private, and the transcript provides only limited information about its financial performance beyond its early revenue and funding announcement.

Collector Cars as an Alternative Investment

• The guest described collectible cars as a possible alternative asset, emphasizing that buyers should start with cars they genuinely like and avoid assuming that every rare or expensive model will appreciate.
  • He said some Ferrari 430 models were beginning to rise in value, while the Ferrari SF90 had declined and might not yet be old enough to attract collectors who grew up with it.
  • He said cars built before 1980 were generally flat or declining, while demand for cars from the 1980s, 1990s, and 2000s was being supported by a new generation of collectors.
  • Rare manual-transmission Lamborghini Murciélago models were cited as an example of cars that proved rarer than many buyers had assumed.
  • The guest also mentioned factory recommissioning programs that can restore and reconfigure some high-end cars, but said their effects on value depend on the model and program.
• Risks discussed included automotive depreciation, concerns about the long-term reliability of some hybrid systems, and the possibility of a market correction. The guest said collector-car borrowing is often limited to around 60%–70% loan-to-value, which may reduce—but does not eliminate—the risk of forced selling.

Takeaways

• Collector cars are an illiquid, specialist asset class, and rarity alone does not guarantee appreciation. The discussion favors careful attention to model history, condition, and the collector audience.
• Do not treat the guest’s examples as broad forecasts: the episode’s comments on particular models reflect his market observations, not guaranteed future returns.