Astromech raises $20M to build a 'biological operating system'
A startup betting that biology can be forecast like the weather just pulled in a $20 million vote of confidence — and a valuation that says investors think the model might actually work.
Astromech, the AI-bio company co-founded by entrepreneur Ben Lamm and geneticist George Church, has raised $20 million in new funding at a $3.8 billion valuation, lifting its total raised to $60 million. The round was led by biotech investor Bob Nelsen and drew in Peak 6, NeoGenesis Capital, Builders VC and CAZ Investments. The pitch is a "predictive model of biology" that doesn't just read today's genomes but projects where living systems are headed tomorrow.
Most computational biology stops at the present tense — here's the genome, here's the disease, here's the protein. Astromech wants the future tense. Its models crunch genomic, evolutionary, biological and functional data to anticipate how organisms will change: genetic bottlenecks, disease progression, drug resistance, the way species react to a shifting environment. Lamm, who also founded the de-extinction lab Colossal Biosciences, frames it as meteorology. "We are building an algorithmic prediction solution," he told Inc. "Think of it like the weather, a complex system that humanity can predict due to specific technology and datasets. We are building the same thing for biology with evolutionary data." Astromech was spun out of Colossal and leans on 3.8 billion years of evolutionary history as a core training signal, blending data from living and extinct organisms with ancestry and biological responses.
Under the hood are two engines working in tandem: one uses deep learning to spot patterns across species, the other reverse-engineers biological systems backward through time before projecting them forward. Church argues the edge is in ancestral regulatory states rather than proteins alone. "Most of the variations that matter for complex traits, for example morphology and longevity, are regulatory rather than coding," he said, "so reconstructing the ancestral regulatory state, not just the ancestral protein, has crucial explanatory power."
The proving ground is human longevity. The company has already mapped 46 longevity-associated genes across a time-calibrated tree of life, studying how the Asian elephant's cancer-suppression mechanism or the bowhead whale's 200-year lifespan evolved, in search of the genomic and regulatory levers behind aging. Next up: bigger research teams, more species, and pilot projects with health and biosecurity partners applying its vulnerability forecasts to real problems.
What to watch: whether "predictive biology" can move from an elegant metaphor to reproducible results — and how the biosecurity world treats a model explicitly built to forecast biological weak points.
If we can forecast evolution like the weather, who decides what to do with the predictions — and who's liable when the model gets biology wrong? Tell us in the comments.
