Owning both AI labs is the hedge. The bet is everything else.

Capital is learning to hold both ends of the same trade. A fund that already owns OpenAI and Anthropic does not need to predict which lab wins the frontier race. It needs to monetize the world that runs on top of whichever lab does.

Deven Parekh of Insight Partners, the $90 billion backer of both labs and Databricks, made that case at StrictlyVC. The argument: the smart AI money is no longer in choosing the winner. It is in applied industries where models become infrastructure, and the returns are not zero-sum across labs. Healthcare is the working example: predicting a 25% heart-attack risk for a patient walking in for something unrelated, drawn from 50 million records at NYU Langone, where Parekh sits on the board.

Insight runs early-stage, growth, and buyout strategies at once, and Parekh frames that breadth as the whole point: a fund that concentrates on a single frontier bet is implicitly forecasting who wins a race still running. A fund that holds both labs and pours the rest into healthcare, legal, and biotech is collecting rent on the race regardless of outcome. Concentration is a bet on a name. Diversification is a bet on the layer the names will share.

The risk Parekh prints openly: open-source models and non-state actors inherit the same toolchain. He treats that as the price of faster drug discovery. The call is not that AI is safe. It is that AI is already infrastructure, and infrastructure pays the layer above it, not the vendor beneath it.