Aug 22, 2026 · 42 min
AI’s value is spreading beyond frontier labs
Martin Casado on Where the Value Is Going in AI
As AI companies absorb unprecedented capital, the key investment question is whether value concentrates in foundation models or spreads across the stack.
- 1AI’s capital intensity lets small teams convert unusually large funding rounds directly into usage, growth, and targeted capability.
- 2Open-source models, specialized systems, and routing platforms challenge the assumption that frontier labs will control every layer.
- 3Cursor and OpenRouter show why strategic optionality can matter before conventional metrics fully capture a company’s value.
Don't miss
Casado explains why the Cursor–SpaceX combination can unite developer data and software expertise with compute and engineering resources.
The brief
Martin Casado joins Theo Jaffee and Sofia Du to examine whether frontier labs will capture the entire AI stack—or whether new layers can retain strategic value.
AI’s unusual capital intensity changes venture economics: companies can deploy enormous sums productively without immediately creating the hiring and organizational burdens of older businesses.
OpenRouter becomes the case study for model access, routing, and specialization, raising the possibility that choosing among models is itself a major AI problem.
The discussion turns to Cursor’s combination with SpaceX, arguing that software expertise, developer data, compute, and engineering resources can create value conventional metrics miss.
The broader lesson is strategic optionality: founders and investors should avoid zero-sum bets on one layer and look for assets that gain importance as the market expands.
OpenAI
AnthropicListen to the full episode and explore every guest, topic, and moment on PodLume.