a16z Partner Casado on AI: AI is the First Technology to Provide Reliable Returns for a $10 Investment
a16z General Partner Martin Casado stated in an interview that AI is the first technology where a reliable return can be achieved with a $10 investment, whereas previous technologies required hiring engineers and waiting two years to see if it would pay off. Casado led a16z's investments in Cursor and OpenRouter.
The interview coincided with SpaceX's $60 billion acquisition of Cursor and Stripe's agreement to acquire OpenRouter. Casado discussed whether labs will win everything, when subsidies will stop, and why RSI is not the correct term, introducing the concept of "self-catalysis."
He pointed out that more private capital will expand the Total Addressable Market (TAM) rather than just inflate valuations; a16z had discussed fully betting on AI seven years before GPT emerged. The stickiness of models exceeded expectations, and true model routing is a complete AI problem. The iteration speed of Cursor is the fastest he has seen, aside from Elon Musk's companies.
Casado believes that once subsidies stop, marketing has been disrupted by AI, and users can now be acquired for $1; Chinese operators are arbitraging $200 subscription plans, and CMOs are becoming CFOs. He elaborated on both sides of the argument regarding labs winning everything and provided an 80/60 judgment for after supply constraints ease in 2028.
The two transactions demonstrate a balance between strategic value and business quality, with Casado emphasizing that the role of VCs is not to know where to build, but to keep moving forward as long as there are mountains to climb.
At the market mechanism level, capital is flowing from strategic buyers to high-iteration AI applications and infrastructure, with funds concentrating on companies that have rapid product cycles and distribution advantages; events are driven by acquisitions and investment returns, benefiting early investors in Cursor and OpenRouter, while pure lab models face higher competitive pressure.
Source: Public Information
ABAB AI Insight
As the head of infrastructure investments at a16z, Martin Casado has long focused on developer tools and network layers, previously leading bets on Cursor and OpenRouter. The two transactions exited within a week at levels of $60 billion and approximately $8 billion, validating his early judgment on fully betting on AI.
In terms of capital pathways, a16z locks in coding assistants and model routing infrastructure through seed and subsequent rounds, directing resources towards companies with high-frequency iterations and network effects, motivated by capturing the value of AI's shift from models to applications and exchange layers, including deep involvement in Cursor's product cycles and OpenRouter's ecosystem development.
Similar cases can be seen in early venture capital bets on infrastructure like Stripe and OpenSea, which were later strategically acquired, as well as Microsoft's integration path with GitHub; the current AI industry is at a stage where strategic buyers are locking in application and routing layers at high premiums, with the boundaries between labs and applications being redefined.
Essentially, this is about capital concentration: strategic buyers acquire high-growth AI tools at extremely high valuations, with the mechanism of directly integrating rapid iteration capabilities and distribution networks into larger platforms, thereby accelerating TAM expansion while raising the entry barriers for newcomers.
ABAB News · Cognitive Laws
- A $10 investment yielding returns is a true watershed moment for AI.
- Iteration speed determines exit multiples, not the number of papers.
- The day subsidies stop, the CMO officially becomes the CFO.