Biohub invests $500M to build AI models of human cells

Chan Zuckerberg Biohub is investing $500 million within a period of five years in developing artificial intelligence models of human cells, tools that can be used to investigate diseases digitally at unprecedented speeds beyond what laboratory studies can offer.

The non-profit organisation, founded in 2016 by Meta Chief Executive Officer Mark Zuckerberg and doctor Priscilla Chan, revealed the funding programme last week as part of the launch of a plan to freely distribute all data created by the technology to researchers worldwide.

The organisation will allocate $400 million for its own initiatives, while another $100 million will be offered as grants to other researchers. Nvidia has been listed as one of the collaborators for the project.

Biohub's Head of Science Alex Rives said the challenge ahead is fundamentally one of data scale. "To build artificial intelligence that can accurately represent the full complexity of biology, we need orders of magnitude more data than exists today," he said at the announcement.

However, despite being the holder of the world’s largest single-cell datasets and developing computing infrastructure designed specifically for biological research, the organisation admits that such massive investments have to be made on a worldwide scale rather than solely by Biohub itself.

In case the model were sufficiently accurate, it would work as a sort of digital laboratory where scientists could predict how the cell reacts to various stimuli, stresses, diseases, or new drugs without performing any experiments in real life.

Biohub suggests that this would help uncover the underlying mechanisms of some unknown diseases and find their treatment.