GenBio AI, a Palo Alto startup co-founded by Nobel laureate David Baker, unveiled AIDO Cell, an AI “world model” it says can simulate a human cell from DNA to whole-cell behavior and predict its response to drugs. It is an early preview, limited to two cell lines, but points toward testing drugs in silico.

[For more news, click here] 

If scientists ten years ago were told they could simulate a living cell inside a computer, they may not have believed it. For decades, biologists trying to emulate cells have been repeatedly defeated by biology’s sheer complexity. However, with the advances in AI, the ambition to do so was revived and on 18th August, 2026, a Palo Alto startup said it had taken a concrete step toward it. GenBio AI, whose co-founders include Nobel laureate David Baker, unveiled AIDO Cell, which it describes as the first AI system able to simulate a human cell across its full biological hierarchy, from DNA and RNA through proteins to the behavior of the whole cell.

A Cell as a ‘World Model’

AIDO, short for AI-Driven Digital Organism, Cell is built as a “world model”. Where most AI models make a prediction in a single pass, a world model keeps an internal state that a user can perturb midway and then watch the consequences unfold. To put it simply, the system is built to sustain sequences of interventions built on one another, similar to successive experiments in a laboratory, rather than being treated as isolated guesses. The comparison GenBio invites is with AlphaFold, the protein-structure predictor. Where AlphaFold models a single protein, AIDO Cell aims to model how an entire cell behaves.

“What's exciting here isn't that we've solved cellular biology – we haven't, at least not yet,” said Baker, who received the 2024 Nobel Prize in Chemistry for computational protein design. “It's that, for the first time, we have a system capable of simulating a cell across the full hierarchy of biological scales, from DNA to whole-cell behavior, in one place, which lets you interrogate it computationally.”

In one early demonstration, GenBio used the system to model the effect of the cancer drug imatinib on leukemia cells and reported that it reproduced the drug's known mechanism across several levels of biology. AIDO Cell currently supports just two human cell lines, K562 and HepG2, both long-standing laboratory workhorses for blood cancer and liver biology. GenBio evaluates it on an in-house Virtual Cell Benchmark that scores 31 metrics across five task families and says AIDO Cell is the only system to span all five, a claim that has not been independently verified. The company is working with NVIDIA on the underlying compute, and three co-founders set out a roadmap in a recent Nature Medicine article.

In the near term, its main use is as a testing ground aimed at trying out drug candidates on the computer first, so researchers can rule out the weak ones before spending time and money testing them in the lab. GenBio is also positioning the project as relatively open, saying it will offer early access to academic, biotech, and pharmaceutical researchers rather than keeping the model behind closed doors, and it plans more capable versions later this year.

An Early Preview, With Real Caveats

The caveats are considerable, and GenBio makes some of them itself, describing the current release as a preview and an early functional demonstration. Keeping in mind the complexity of the human body, two cell lines is a narrow slice of human biology. The harder problems are well known as keeping small errors from compounding as predictions ripple across biological scales, and proving that what the model predicts actually holds up in living cells rather than only on a benchmark. Those validation questions, not the ambition, will decide how useful virtual cells become.

For now, GenBio is careful not to overclaim, and Baker's framing is the honest one. A virtual cell has not been solved but a system that lets researchers interrogate a simulated cell across every biological scale at once, if it proves accurate, would be a genuinely new instrument for biology.

Related Articles

Amber Lands €7M to Turn Enterprise Knowledge into Autonomous AI

How 1001 Raised $30 Million to Bring Sovereign AI to Ports, Airports and Energy Grids

Cequence Security Launches Agent Personas to Automate AI Agent Governance Across the Enterprise