Written by AI.How we work
Topic
Machine-learning models trained on biological substrates such as living human tissue rather than animal models, cell cultures or organoids.
Meta, Google DeepMind and Isomorphic Labs put $300 million into Biohub's $1.8 billion Virtual Biology Initiative and get the data before the public does. Startups that sell proprietary cell data will now be priced against a free dataset that their best-funded rivals see first.
Reality
The $20mn round buys roughly 0.53 per cent of the company, and no product exists yet. Read the number as a reference price for pre-revenue bio-AI, not a verdict on the science.
Reality
The claim arrived in a founder profile with no endpoint, no sample count and no outside lab. Its nearest customer is the cosmetics brand that shares a scientist with the company.
Reality
Dario Amodei says Claude read the literature and genome data, found something and proposed the experiments, while Anthropic's scientists chose the topic and ran them. He says the finding's function is not yet clear.
Reality
The AlphaGenome Atlas replaces a compute-heavy programming interface with a browser lookup for every possible single-base change in human DNA. The evidence that its impact score flags disease-linked mutations sits in two preprints.
Reality
TranscriptFormer learned gene expression the way a language model learns text, across 12 species from single-celled yeast to humans. Its cross-species matches arrive as hypotheses a bench experiment still has to test.
Reality
AlphaGenome Atlas turns per-variant inference into a table lookup and adds a ranking score, which changes what a genomics budget buys this quarter well before it changes what a clinic can act on.
Reality
The spinout's valuation is 63 times every dollar of capital it has ever taken, on datasets gathered for de-extinction work. Anyone holding irreproducible data has an unpriced clause in a contract.
Reality