Amazon Starts Selling AI Models Built From Rat Brain Cells

Starting Tuesday, select Amazon Web Services customers get access to an AI video model derived from rat brain cells, as part of a limited preview. The offering marks the first mainstream cloud distribution for The Biological Computing Company (TBC), a Baltimore startup that translates patterns from living neurons into software — and which has quietly raised more than $50 million, including a previously unreported $25 million round.

AWS and TBC say they expect the model to roll out to all AWS enterprise customers soon. The deal matters because it puts biological computing — a once-fringe field that runs actual wet labs alongside computer labs — inside the distribution channel of the world's largest cloud provider, at a moment when AI companies are scrambling to cut the cost of inference.

"We figured out a way to code information, like images for example, to the biological material," says TBC cofounder Alexander Ksendzovsky, now the company's CEO. "We then observe how the biology processes that information, and then we build a tool that mimics that process."

How the technology works

TBC was founded four years ago in Baltimore by two neuroscientists and neurosurgeons, Ksendzovsky and Jon Pomeraniec, who serves as president and COO. The company now has about 35 employees, some of whom work directly with rat brain cells and human stem cells in a small research lab it opened last year in San Francisco.

The living cells sit on multi-electrode silicon arrays made by 3Brain, a Swiss biotech company. Researchers send electrical stimulation patterns into the neurons, record the responses, and mine that neural activity for computational patterns that can be folded into existing AI models — specifically video generation models.

The video focus was both scientific and practical. The physical layout of the electrode arrays means visual information maps onto the grid more readily than text or language. From there, a business case emerged: model how neurons process images, and use those insights to make video generation AI more efficient.

Pomeraniec credits Jeff Dean, the prominent AI researcher and TBC investor, with pushing the company toward video first. Because established industry benchmarks exist for video models, TBC could test its technology on already-solved problems to prove it represented a real scientific advance. "Jeff said if we could do that, there's the promise of doing more complex, interesting things with our AI models down the line," Pomeraniec says.

The funding picture and the performance claims

TBC raised its first significant round earlier this year — $25 million led by Primary Venture Partners — and closed in March. Shortly after, it secured another $25 million, bringing total funding above $50 million. The second round has not been previously reported.

The startup claims its model, measured against the open-source frontier video generation model it runs on, is up to five times faster at video generation and significantly lowers inference costs. The company declines to name the comparison model. Previously, TBC's technology was only available through a neocloud provider called Bluesky Compute.

A pragmatic bet in a hard field

This is not Amazon's first biologically derived computing offering. AWS also works with Cortical Labs, an Australian company that combines lab-grown neurons with silicon chips and sells "wetware as a service," plus a multi-thousand-dollar biological computer that it says can keep neurons alive for six months.

Deap Ubhi, global director of technology for startups at AWS, says TBC's appeal lies in its pragmatism. Rather than reinventing the transformer, "they're working within existing standards of the generative AI space and saying, 'How can we make the current visual models more efficient?'" he says.

He is also candid about the open question: scale. "I think with any model like this, the one thing you have to ask is, if you push it to the extremes, will you still see improvements? If you have customers who want to generate longer-form videos, ten minutes, an hour, what's the potential loss in fidelity over the course of time?" Ubhi asks. As new AWS customers test the model, TBC's next phase will show whether biology-derived efficiency gains survive contact with production workloads.