Valuation and Market Reality

• Anthropic is reportedly targeting a $2T valuation, a figure that fails standard discounted-cash-flow models even under heroic, no-cost, perpetual-revenue assumptions.0:21
• Bankers frequently bridge this valuation gap by citing rapidly inflating Total Addressable Market (TAM) claims, a technique historically used to justify Amazon and SpaceX but which rarely maps to realized earnings.9:30
• The sector faces a double-edged sword from high interest rates, which simultaneously increase the cost of building data centers and decrease the present value of the long-term profits labs promise.3:14

The Financing Loop

• SoftBank is heavily entangled in the ecosystem, issuing record-setting 9.75% junk bonds to fund OpenAI while OpenAI rents space from SB Energy—a developer that has yet to turn on a single data center.18:55
• Nvidia acts as the core guarantor, pledging up to $105B for campuses in exchange for exclusive hardware use, while simultaneously holding significant equity stakes in the very labs that buy its chips.19:48
• Large tech firms like Amazon and Google are reporting significant portions of their net income from unrealized gains on their private stakes in these AI startups, making their own valuations partially dependent on the labs' private marks.21:01

Competitive Pressures

• Nvidia currently trades at less than 17 times forward earnings, historically cheap for a growth leader, reflecting investor fears of peak-cycle margins and the risk that AI spending may eventually slow.22:10
• AI model prices are collapsing, with costs per performance level falling 13-fold annually, while labs face intense pressure from open-weight competitors and customers who now route tasks between models to slash bills by up to 40%.26:42