i
News
News · 2026-10-07

Biohub leads a $1.8 billion effort to predict cell behavior

@neuronium_ai @neuronium_ai

Biohub, the nonprofit founded by Mark Zuckerberg, is leading a $1.8 billion program to build AI models that predict how cells behave. The effort will produce biological datasets for training those models, with different access rules for commercial and government-funded work. Commercial sponsors get exclusive access to the data they pay for during the first year; afterward, it becomes public. The first dataset is expected in about a year.

Cover: Biohub leads a $1.8 billion effort to predict cell behavior

Who gets the data first

Alex Reeves, Biohub’s head of scientific work, said commercially funded datasets will be restricted to their sponsors for one year before becoming publicly available. That restriction will not apply to data funded by the government.

The first dataset is expected in roughly a year, leaving a long gap between the program’s announcement and its first planned output. The announcement does not say how many datasets the $1.8 billion will fund or how the program will measure whether its models predict cell behavior accurately.

A crowded field, an open question

Other AI companies are also investing in biology. Anthropic has built its own biology lab to develop drugs with AI, while the OpenAI Foundation is directing more than $125 million toward biological and medical datasets.

I think the one-year exclusivity period is the program’s defining trade-off: it gives commercial sponsors a reason to pay for data while delaying wider access. The harder question is whether a year is enough to make the investment worthwhile for sponsors without slowing the public value the program promises. That balance will matter as much as the models themselves.

Daily AI news

Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.

Only what matters — every day

Follow on X