The Biggest IPO in History Is Undervalued

SpaceX (ticker: HQ) is presented as a buy-ahead-of-IPO opportunity driven by Starlink expansion and plans for space-based AI data centers that could host GPUs and serve major AI customers.

The podcast argues SpaceX is undervalued ahead of its IPO, citing: Starlink’s explosive growth trajectory; plans to deploy space-based AI data centers (satellites designed to host NVIDIA GB300-class GPUs and serve multiple AI vendors); a potential hardware-scaling advantage that could be agnostic to chip vendor (NVIDIA, Google, Amazon); and the strategic optionality from partnerships with large AI players and cloud providers. The hosts frame the thesis as long-term and low-time-preference, while noting this is opinion content and not financial advice.

Discussion of Leopold Aschenbrenner’s AI-focused portfolio, including a reported NVIDIA short and a private investment in Anthropic. Covers NVIDIA’s bond offering and a shift in the AI infrastructure trade toward power, networking, and data-center buildout rather than just chips.

Claims that Anthropic released then withdrew frontier models after a U.S. export-control/national-security notice. Amazon Web Services allegedly revoked access due to verification and jailbreak concerns. The takeaway: rising government constraint on frontier AI distribution, increased compliance burden, and potential advantage to large regulated platforms and defense/compliance vendors.

Episode focused on SpaceX as a potential buy ahead of its IPO. Topics include valuation, Starlink growth, space-based AI data-center plans, OpenAI data-center reports, Anthropic’s Fable 5, and Apple WWDC AI updates. The hosts argue SpaceX could serve multiple AI customers with satellites configured for significant GPU capacity.

Thematic discussion of 'AI loops'—longer-running, more autonomous AI workflows—and the implications for compute, cost, and the continuing role of human judgment. Relevant as background on why new infrastructure modalities (including space-based compute) might be valuable.

Coverage of Anthropic’s model release emphasizing stronger capabilities constrained by safety controls and compute/pricing limits. Illustrates the tension between frontier model capability and regulated or capped access—context for infrastructure demand and compliance needs.

Recap of Apple WWDC AI updates, noting Apple’s on-device plus encrypted cloud approach and potential EU rollout complications. Relevant to hardware/edge AI narratives that overlap with infrastructure and privacy considerations.

Podcast commentary on Microsoft and NVIDIA announcements: Microsoft’s Copilot/agent adoption questions, NVIDIA hardware momentum, and cloud compute constraints. Useful context for competitive dynamics among major cloud and chip providers.

Report of an alleged exploit abusing Meta’s AI-driven account recovery to hijack accounts. Emphasizes AI security risks and the likelihood of increased security spending and regulatory scrutiny—factors relevant to enterprise demand for secure, verifiable infrastructure.