• Consider CoreWeave (CRWV) on weakness as a higher-risk AI infrastructure opportunity; its $104 billion backlog supports demand, but verify it can secure GPUs, electricity, and financing to deliver.
• For broader AI exposure, favor a diversified approach to AI developer tools and model platforms rather than betting on a single private company; assess valuations carefully, as several cited companies already carry lofty private-market valuations.
• Treat claims about SpaceX’s rapid compute rollout and projected $500 billion–$1 trillion revenue run rate by end-2027 as speculative, not a verified investment case.
• Monitor Science Corporation and Neuralink, but wait for evidence of scalable adoption—especially non-invasive products—before treating them as actionable investments.

CoreWeave (CRWV)

• The discussion was bullish on CoreWeave and described its $104 billion backlog as evidence of strong demand for AI computing capacity.
  • One speaker argued that the market’s negative reaction to CoreWeave and other AI infrastructure companies could be a buying opportunity, saying demand is outpacing the company’s ability to deliver.
  • The speakers also noted concerns about rising bond yields and the difficulty of financing GPU clusters, energy systems, and cooling infrastructure.
  • They suggested that Erebor, a new bank mentioned in the episode, could help if it becomes better at valuing GPU clusters and related infrastructure as loan collateral.
• SpaceX was described as able to build compute capacity faster than CoreWeave, but the speakers argued that CoreWeave could still benefit even if competitors are more efficient.

Takeaways

• The investment case presented depends on CoreWeave turning its backlog into delivered capacity and revenue. Watch whether it can secure the GPUs, electricity, and financing needed to meet demand.
• The speakers saw the stock’s weakness as a potential opportunity, but the transcript also highlights financing costs and execution capacity as important factors.

SpaceX

• The speakers described SpaceX as a major potential competitor in AI data-center capacity, saying it could bring compute online faster than other providers.
  • They cited a claimed six-month return-on-investment period for new infrastructure and a forecast of $500 billion to $1 trillion in revenue run rate by the end of 2027.
  • The episode also said SpaceX was trading at $165, above its IPO price. The transcript’s comments about its IPO and valuation are unclear, so these figures should be treated as claims made in the discussion rather than independently verified details.
• One speaker called SpaceX the strongest play for rapidly building compute capacity, while the other argued that CoreWeave could still succeed even if SpaceX is more efficient.

Takeaways

• The opportunity described is exposure to AI data-center growth, but the revenue and payback claims are highly ambitious. The discussion itself emphasizes how much the outcome depends on building capacity and capturing customer demand.

Science Corporation

• Science Corporation was discussed as a Neuralink competitor, founded by Neuralink co-founder Max Hodak.
• The speakers were enthusiastic about the potential to help people with vision loss and other neurological conditions, but questioned whether those medical applications alone provide a large enough market for venture-scale investment returns.
• They saw a potentially much larger opportunity if the company can develop non-invasive wearables that read or respond to brain signals without surgery.
  • The discussion compared adoption prospects with familiar devices such as earbuds and smartwatches.
  • It also mentioned that similar products are being sold in China, but did not provide details about Science Corporation’s own sales.

Takeaways

• The investment thesis described is conditional: medical applications may have significant social value, while a broad non-invasive consumer product could expand the commercial opportunity.
• The speakers said they would want to see the technology develop and begin selling before drawing stronger conclusions. No valuation or investment recommendation was given.

Neuralink

• Neuralink was discussed alongside Science Corporation as part of the brain-computer interface sector.
• The speakers questioned the practical scalability of surgically implanting chips in very large numbers of people and raised concerns about whether consumers would choose invasive implants.
• They suggested that widespread adoption could become more plausible if brain-computer interfaces become socially or professionally advantageous, but presented this as speculation.

Takeaways

• The discussion points to adoption and scalability as key questions for Neuralink and similar companies.
• The speakers viewed non-invasive devices as a potentially easier route to a broader market than brain implants.

ElevenLabs

• ElevenLabs was described as an AI company focused on voice, with the episode citing a $22 billion valuation in September.
• The speakers pointed to several potential applications:
  • AI-powered customer service and real-time voice conversations.
  • Voice cloning and tools for editing speech, such as correcting a stumble in a recording.
  • Creating original music and sound for advertising, videos, and other content.
  • Combining generated research, scripts, voice, and video into automated content workflows.
• One speaker said a newer voice model was intended to make conversations feel more natural, with less delay and better voice inflection.

Takeaways

• The investment case discussed rests on ElevenLabs continuing to improve voice technology and converting applications such as customer service and content production into sustained use.
• The $22 billion valuation was cited without further valuation analysis or a specific recommendation.

Modal

• Modal was described as an AI infrastructure and developer-tools company offering services for model inference and training.
• The episode cited $1 billion in annualized run-rate revenue, a $750 million funding round, and a $15.75 billion valuation.
• The speakers said businesses may use platforms like Modal to customize models and work with open-source AI, rather than relying solely on expensive offerings from companies such as Anthropic and OpenAI.
• One speaker suggested that Modal and similar businesses could have significant long-term potential, while noting that the key question is how large these companies can ultimately become.

Takeaways

• The discussion favors the broader AI developer-tools and model-platform opportunity, but investors would need to assess whether rapid revenue growth can support high private-market valuations.
• The episode did not provide a specific recommendation for Modal.

Base10, Together AI, and Fireworks AI

• These companies were named as peers or related opportunities in the AI infrastructure and developer-tools space.
• One speaker said an investor seeking exposure to the sector might consider allocating across multiple companies, including Modal, Base10, Together AI, and Fireworks AI.
• The rationale was that businesses building, customizing, and serving AI models may benefit as companies seek alternatives to relying exclusively on large, costly model providers.

Takeaways

• The discussion favors broad exposure across AI platforms and developer tools rather than relying on a single company.
• No valuations or specific recommendations were given for Base10, Together AI, or Fireworks AI.

Databricks

• Databricks was cited as being valued at more than $200 billion in the secondary market.
• The speaker used that valuation as evidence that AI infrastructure and developer-tools companies could potentially grow to very large valuations.

Takeaways

• Databricks served as a comparison point for the possible scale of AI-related private companies, not as a specific buy recommendation.
• The transcript did not discuss Databricks’ financial performance or a suitable entry valuation.

AI infrastructure and model platforms

• The speakers expressed a broadly bullish view of AI infrastructure, arguing that demand for computing capacity is strong and that many investors may not yet appreciate how widely AI can be used.
• They described AI as being in its “first inning” and argued that the opportunity extends beyond data centers and GPUs to developer tools and platforms that help businesses customize models.
• They also identified constraints: companies may struggle to obtain enough GPUs and electricity, and providers may not be able to deliver capacity quickly enough.

Takeaways

• The investment theme discussed spans data-center capacity, cloud computing, and tools for building or adapting AI models.
• The speakers’ positive outlook depends on AI adoption continuing to grow and companies converting demand into delivered services and revenue.