Rat brains power new AI

Amazon Web Services is now offering AI models built from rat brain cell patterns through The Biological Computing Company, starting with a limited preview for select customers.

How the tech actually works

The startup encodes images onto rat brain cells, observes how the biology processes them, then builds software tools that mimic those neural patterns.

“We figured out a way to code information, like images for example, to the biological material. We then observe how the biology processes that information, and then we build a tool that mimics that process.”

Not Amazon's first bio bet

Amazon already works with Cortical Labs, an Australian company that combines lab-grown neurons with silicon chips and sells products under the name wetware as a service.

Why Amazon chose this startup

The Biological Computing Company takes a pragmatic approach by working within existing AI standards to make video models more efficient, rather than trying to reinvent the entire system.

Biological computing gains momentum

As AI companies race to make models more efficient, the once-fringe field of biological computing is gaining traction, though it still faces unique challenges keeping cells alive and translating biology into code.

From Baltimore to San Francisco

Founded four years ago by neuroscientists Alexander Ksendzovsky and Jon Pomeraniec, the startup has raised over fifty million dollars and now employs about thirty-five people across two labs.

Why video generation first

The company chose to focus on video because the physical layout of their multi-electrode arrays made it natural to map visual information onto the grid, and video generation has established industry benchmarks to measure success.

Performance claims and limits

The startup claims its model is up to five times faster at video generation and significantly lowers inference costs compared to open-source alternatives, but scaling to longer videos remains an open question.

The scaling challenge ahead

As Amazon distributes the technology to enterprise customers, the real test will be whether it maintains consistency and quality when generating longer videos or handling more complex tasks.