• Consider AI infrastructure as a diversified theme across data centers, inference providers, and semiconductors, while comparing valuations and business fundamentals rather than chasing momentum.
• Among the private inference companies discussed, Together AI was viewed as a potentially attractive entry below $10 billion, but verify its revenue and growth comparability with peers.
• NVIDIA (NVDA) was highlighted as a possible buy during a market downturn; the $10 trillion market-cap idea is a long-term opinion, not a timed price target.
• Be cautious about paying elevated valuations for Baseten, FAL, or Fireworks AI; the cited valuations have risen sharply, and the discussion offered no evidence that revenue supports them.

AI Infrastructure

• The speakers see AI infrastructure as an attractive investment theme and say they would overweight it relative to other areas.
• They divide the space into data centers, AI inference providers, and semiconductor chip providers. Their suggested approach was to hold exposure across all three, initially with roughly equal weights and possibly a tilt toward inference providers.
• They favored diversification across multiple companies rather than choosing only one or two. They also stressed comparing entry valuations: a strong company may still be a less appealing investment if its valuation is already very high.
• The discussion flagged that some AI companies’ valuations may be difficult to justify without substantial future revenue growth. The speakers also noted that technical complexity can make these businesses hard to evaluate.

Takeaways

• The discussion supports considering AI infrastructure as a broad theme, not treating every company in it as interchangeable.
• Compare companies within their category, and weigh valuation against revenue, product maturity, and the capital needed to grow. The speakers’ views are opinions, not guarantees of future performance.

VAST Data

• VAST Data was described as a software-layer company focused on maximizing memory and storage capacity, rather than a hardware maker.
• The episode cited $10 billion in customer commitments for infrastructure in 2026: 14 customers had committed, including four with commitments above $1 billion each.
• One speaker viewed the commitments as evidence that the company was becoming difficult to ignore, while acknowledging that the underlying technology—such as VRAM and flash storage—is technically difficult to assess.
• The speakers were uncertain whether VAST should be grouped with data centers or semiconductors, but discussed it as part of the broader AI infrastructure stack.

Takeaways

• The reported commitments may indicate substantial customer demand, but commitments alone do not establish the company’s long-term economics or investment value.
• Further evaluation would need to clarify what the commitments mean in practice and how VAST’s software differentiates itself. The speakers specifically noted the challenge of evaluating the technology in depth.

Baseten

• Baseten was discussed as an AI inference provider. The speakers cited a valuation of about $13 billion in June, followed by a reported financing process involving a valuation of $26 billion and a subsequent raise at around $30 billion.
• One speaker said they liked the company, but also discussed the importance of comparing its valuation with peers and with its revenue.

Takeaways

• Baseten is one example of the rapidly rising valuations in AI inference.
• The discussion suggests comparing its growth and revenue with similarly positioned companies before judging whether its valuation is attractive.

fal.ai (FAL)

• FAL was described as providing inference for video, audio, and images.
• The speakers cited a valuation of $8 billion in April and $16 billion later, describing it as having doubled.
• One speaker found the video-focused business especially exciting, while also recommending diversification across companies in the space.

Takeaways

• FAL’s focus on media-related AI inference was a source of enthusiasm in the discussion.
• Its reported valuation increase makes it important to consider whether the company’s future growth can support its current price.

Fireworks AI

• Fireworks was described as an AI inference provider. The speakers said market chatter suggested a possible financing round at a valuation of $30 billion to $40 billion.
• They characterized the increase from its earlier valuation as roughly a 2x or greater move, but did not provide an earlier valuation figure.

Takeaways

• The reported potential valuation is a market indication, not a confirmed financing outcome.
• As with other inference providers, the discussion points to comparing valuation with revenue and business progress rather than relying on momentum alone.

Together AI

• Together AI was described as an AI inference provider.
• One speaker favored it because it was valued below $10 billion, while some peers were valued at roughly $15 billion to $40 billion.
• The speaker said Together AI’s revenue was “on par” with those peers and called its valuation an attractive entry point.

Takeaways

• The comparison raised a possible valuation gap relative to peers, based on the speaker’s view of revenue.
• Investors would need to verify the revenue comparison and assess whether the companies have comparable growth prospects and business models.

Etched

• Etched was discussed as a semiconductor-related company with a valuation of roughly $20 billion to $30 billion.
• The speakers noted that it did not yet have a product out, raising questions about how its valuation could be justified.
• They also discussed the possibility that future funding needs and venture-capital dilution expectations can influence high private-market valuations.

Takeaways

• The lack of a product was a specific concern in the discussion, especially alongside a high valuation.
• The speakers’ comments underscore the importance of assessing execution and future financing needs, not just the appeal of a company or its founders.

Positron

• Positron was described as a semiconductor-related company valued at around $6 billion to $7 billion.
• One speaker said they might prefer that entry valuation to Etched’s higher valuation, while allowing that Etched could be the better company.

Takeaways

• The comparison illustrates the speakers’ emphasis on entry price: a lower valuation may be more appealing, but it does not by itself establish that a company is a better investment.
• The discussion provided no further details about Positron’s product, revenue, or prospects.

Micron and SanDisk

• Micron and SanDisk were cited as public companies whose stocks had performed strongly over the preceding year and a half, amid attention on memory chips and hardware.
• The speakers contrasted that hardware interest with VAST Data’s software-layer approach.

Takeaways

• The discussion highlighted strong past performance in memory-related stocks, but did not offer a price target or a specific buy recommendation.
• Past stock performance does not establish future returns.

Revolut

• Revolut was discussed at a reported valuation of $115 billion.
• One speaker questioned whether its valuation was justified compared with SoFi, which was cited at a market capitalization of about $21 billion. They said they would need to examine Revolut’s revenue to assess the comparison.
• The speakers also questioned how much growth remained in fintech and whether Revolut could eventually reach the scale of a major bank. One noted Revolut’s international approach, including moving operations from the UK to Dubai, as a possible factor in its growth.
• A potential future public listing was mentioned as a possibility, not a confirmed plan.

Takeaways

• The key issue raised was whether Revolut’s revenue and growth prospects support its much higher valuation than SoFi’s.
• The discussion expressed uncertainty about the remaining growth opportunity in fintech and did not provide a specific recommendation.

SoFi

• SoFi was cited as having a market capitalization of about $21 billion and used as a public-market comparison for Revolut’s $115 billion valuation.

Takeaways

• The comparison was presented as a prompt to examine the companies’ relative revenue and valuation, not as a judgment that either stock was mispriced.
• No specific SoFi price target or recommendation was given.

Stripe

• Stripe was cited at a secondary-market valuation of about $190 billion.
• One speaker was positive on the company’s prospects, arguing that AI agents could create new payment activity and that Stripe could provide payment rails for those transactions.
• The discussion also mentioned Stripe’s work on payment infrastructure and tokenized payments, while noting uncertainty about regulation.
• The speakers debated how much further a company already valued at this scale could grow, with one suggesting a possible path to $500 billion as an illustration—not a stated price target.

Takeaways

• The investment thesis discussed was that AI agents could create additional demand for payment services and benefit Stripe.
• That thesis depends on AI-agent commerce becoming meaningful and Stripe capturing the resulting activity; the discussion also identified regulatory uncertainty around tokenized payments.

NVIDIA (NVDA)

• One speaker said they believed NVIDIA could reach a $10 trillion market capitalization and suggested the current market downturn might be a good time to buy.
• No timeline for that market-cap figure was given.

Takeaways

• This was a bullish opinion, not a specific near-term forecast or a guaranteed outcome.
• The broader discussion emphasized comparing public-market opportunities with private-company investments based on expected returns and risk.

JPMorgan Chase (JPM)

• JPMorgan Chase was cited as having a market capitalization of about $900 billion.
• It served as a scale comparison when the speakers questioned whether Revolut could grow enough to justify its valuation and eventually become a giant financial institution.

Takeaways

• The comparison highlights the scale Revolut would need to reach to approach a major incumbent bank.
• No investment recommendation or forecast for JPMorgan Chase was provided.

Anthropic, OpenAI, and Anduril

• The speakers mentioned potential major-company public-market events: Anthropic was said to be conducting a roadshow, OpenAI was described as likely to go public “next year,” and something might happen with Anduril the following year.
• These were presented as expectations, not confirmed listing plans. The episode did not specify the calendar year for “next year.”

Takeaways

• The comments point to possible future public-market access to large private companies, but the timing and outcomes were uncertain.
• The discussion offered no valuations, price targets, or specific recommendations for these companies.

AI Agents and Payment Infrastructure

• The speakers discussed AI agents as a potential source of new online transactions, which could create demand for payment providers such as Stripe.
• They also noted that tokenized payments could be an additional opportunity, while acknowledging uncertainty about regulation.

Takeaways

• The theme is a possible long-term growth driver for payment businesses, rather than a standalone investment recommendation.
• The speakers’ stated risk factor was regulatory uncertainty around tokenized payments.