• Treat NVDA as a watchlist holding rather than making a headline-driven trade: allegations remain unproven, and the board’s investigation and any export-control developments are key near-term catalysts.
• Be cautious with data-center developers and infrastructure stocks: permitting, community opposition, and power availability could delay planned projects, especially those targeting 2028–2029.
• Track GOOGL for evidence that newer models and enterprise products can reverse its reported decline in enterprise LLM API spending; the survey alone is not a definitive sell signal.
• Favor software companies that can demonstrate customer value while controlling AI costs, as spending is growing rapidly but the discussion provides no clear evidence that AI is broadly displacing core enterprise platforms.

NVIDIA (NVDA)

• The episode discussed allegations that CEO Jensen Huang had a relationship with an executive at Singapore-based Megaspeed, which U.S. officials reportedly suspect may be connected to Chinese intelligence. These remain allegations; NVIDIA’s board said it had no current reason to believe national security or company secrets were compromised.
• NVIDIA’s board opened an independent investigation, and Huang apologized to employees for being unfaithful to his wife. The discussion emphasized Huang’s importance to NVIDIA’s leadership and his role in U.S.-China chip policy.
• The episode reported that NVIDIA shares initially fell about 8%, then were down about 3.7% at the time of recording, while several other major technology stocks were higher. This was a brief market snapshot, not a price target.
• Risks raised included possible congressional scrutiny, shareholder litigation, export-control concerns, governance and key-person risk, and potential damage to Huang’s influence in Washington. The transcript also noted that the board had not announced a leadership change.

Takeaways

• Treat the situation as a material governance and policy uncertainty, but distinguish reported allegations from established findings.
• Monitor the board investigation and any official developments around export controls or congressional scrutiny. The episode did not make a specific buy-or-sell recommendation.

Data Centers and AI Infrastructure

• Voters backed limits on data centers in 40 of 41 local ballot measures discussed in the episode. Candidates who supported slowing or pausing development also won prominent state and Senate races.
• The concerns cited included electricity bills, water use, local control, and who pays for infrastructure. The episode said the results could slow planning for new projects, especially those expected to come online in 2028–2029; construction already underway is less likely to stop.
• One estimate cited suggested 25%–35% of the following year’s U.S. data-center build-out could be at risk. Another article suggested nearly two-thirds of planned 2028 capacity could face delays. These were estimates reported in the discussion, not confirmed cancellations.
• OpenAI proposed “compute campuses” and a community compact that would commit the company to paying for energy and infrastructure, protecting local resources, and creating local jobs. The hosts viewed this as a response to public opposition, not proof that permitting challenges are resolved.

Takeaways

• For investors exposed to data-center developers and related infrastructure, permitting, community opposition, power availability, and project timing are key risks to assess.
• The discussion suggests slower build-outs and higher community obligations may be more likely than a complete halt to AI infrastructure investment.

Alphabet (GOOGL, GOOG)

• Menlo Ventures’ survey of nearly 400 U.S. enterprise decision-makers estimated that Google’s share of enterprise LLM API spending fell from 21% to 9% year over year. The hosts noted that newer Google models could change the picture.
• The same survey estimated that Anthropic, OpenAI, and Google together accounted for 69% of AI deployment spending, up from 50% the previous year.
• Google was also described as developing AI tools for schools and as opening Siri-like AI actions to developers through Apple’s software ecosystem; the latter is an Apple development, not a Google product.

Takeaways

• The survey points to increased competitive pressure on Google in enterprise AI, but it is a limited sample and a snapshot in a fast-moving market.
• Track whether Google’s newer models and enterprise products regain adoption rather than assuming the reported spending-share decline will persist.

OpenAI (Private)

• OpenAI was described as one of the leading AI providers. In Menlo Ventures’ survey, OpenAI and Anthropic together accounted for 86% of enterprise LLM API spending, up from 67% the previous year.
• OpenAI outlined a proposal for future compute campuses that would give communities a say in development and commit the company to paying its own energy and infrastructure costs.
• The hosts suggested that data-center constraints could eventually lead AI providers to raise prices or limit usage, particularly for compute-intensive features such as video generation and agents. This was a forecast, not an announced OpenAI pricing change.

Takeaways

• OpenAI’s enterprise position appears strong in the cited survey, but infrastructure availability, regulation, and the cost of compute could affect its growth and pricing.
• OpenAI is not publicly traded, so the episode did not identify a direct public-stock investment.

Anthropic (Private)

• Menlo Ventures’ survey estimated Anthropic’s share of enterprise LLM API spending rose from 40% to 55% year over year.
• METR’s assessment found that AI is accelerating Anthropic’s AI research; the hosts summarized the report as indicating roughly two years of progress in one year. METR did not find public proof of AI independently creating successive generations of more powerful models.
• Much of the evidence in METR’s public report was redacted at Anthropic’s request, making some conclusions difficult for outsiders to verify. An Anthropic co-founder said many of METR’s conclusions matched his own.
• The episode noted that Anthropic’s research acceleration could intensify competitive and societal disruption, while also emphasizing that human researchers remain involved.

Takeaways

• Anthropic’s reported enterprise momentum and research progress are notable, but the limited public evidence and redactions make independent assessment difficult.
• Anthropic is privately held; no direct public-stock investment was identified.

AI Software, SaaS, and Enterprise Applications

• Menlo Ventures estimated that enterprise AI spending reached an annualized $177 billion, up from $33 billion the prior year. The report also found that companies were spending substantially more than they had planned.
• Coding tools and enterprise chatbots accounted for 76% of AI application spending in the surveyed group. Menlo said it found limited evidence that companies were replacing core systems such as Salesforce or Workday with AI.
• LinkedIn reportedly replaced some third-party AI models with smaller models it developed in-house to keep costs down as usage increased.
• HubSpot was reported to be considering a “headless” approach in which customers use AI tools such as ChatGPT or Claude to work with HubSpot data. It was also testing pricing based partly on usage or work performed.
• Zapier’s CEO argued that combining AI agents for tasks requiring judgment with conventional automation for repeatable steps can reduce costs. The episode treated this as a useful cost-control approach, while noting that Zapier has a commercial interest in the claim.

Takeaways

• AI spending growth creates opportunities for software providers, but it also raises questions about pricing, margins, and whether customers will interact with software through AI rather than its own interface.
• Look for companies that can show clear customer value and manage model costs. The episode did not establish that core enterprise software is already being broadly displaced.

Apple (AAPL)

• Apple opened Siri AI capabilities to developers in its latest operating-system betas, enabling apps to expose actions such as booking reservations. A broader customer rollout was described as planned for later.
• The hosts said Siri had improved in their testing and that Apple appeared to be focusing more on personal agents, while noting that the product was not yet equivalent to computer-use agents.

Takeaways

• Apple’s developer access and planned Siri updates point to an effort to compete in AI-enabled personal assistance.
• The episode provided no adoption, revenue, or financial-impact data, so the product update alone does not establish the scale of the investment opportunity.

Thinking Machines Lab (Private)

• The startup led by former OpenAI CTO Mira Murati was reported to be nearing a $5 billion funding round led by Accel at a $40 billion valuation.

Takeaways

• The reported valuation reflects strong private-market interest in AI, but the episode did not provide operating results or evidence that the valuation is supported by current revenue.
• This is a private-company opportunity, not a publicly traded stock.

Amazon (AMZN), Meta (META), and SpaceX (Private)

• At the time of recording, the hosts said Amazon and Meta shares were each up about 5% that day, while NVIDIA was lower. They also said SpaceX was up about 5%; SpaceX is private.
• These figures were short-term market observations and were not accompanied by company-specific investment analysis.

Takeaways

• The episode offered no substantive investment thesis for Amazon or Meta based on this comparison; the reported one-day moves should not be treated as a signal.
• SpaceX was mentioned as a private company, including in connection with Elon Musk’s idea of placing data centers in orbit. The discussion did not assess the feasibility or investment case for that concept.