Midterms 2026See who we think should earn your vote, based on our standardsThe guide →
WRITTEN IN PLAIN AMERICAN ENGLISH.
CLAY TRIBUNE.
Advertisement

Amazon Lets Customers Rent AI Models Built From Rat Brain Cells

Amazon lets customers rent AI models built from rat brain cells, per startup TBC. A look at how wetware computing works and what it promises.

By mitch·4 min read
A petri dish holding rat brain tissue connected to a microelectrode array sits beside a glowing server rack in a laboratory.

Amazon is letting customers rent AI models built from rat brain cells, and the startup behind them wants you to know it is not trying to reinvent transformers.

Starting Tuesday, select Amazon Web Services customers can access The Biological Computing Company’s “rat brain” AI model as part of a limited preview. The startup, which goes by the acronym TBC, says the tech is aimed at video generation. Amazon expects the offering to expand to all AWS enterprise customers soon.

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

Advertisement

The Brain Cells Behind the Models

TBC was founded four years ago in Baltimore, Maryland by two neuroscientists and neurosurgeons: Ksendzovsky, now CEO, and Jon Pomeraniec, president and COO. The company raised $25 million earlier this year from a group led by Primary Venture Partners, its first significant round of funding. Shortly after, it secured an additional $25 million, bringing its total to more than $50 million.

In San Francisco, TBC runs a small research and development lab where employees work directly with rat brain cells and human stem cells. The living cells sit on multi-electrode silicon arrays made by 3Brain, a Swiss biotech company. Researchers send electrical stimulation patterns into the neurons and record their responses. TBC then translates that neural activity into software to improve existing AI models.

Why Video First

The decision to focus on video was both scientific and practical, says Ksendzovsky. The multi-electrode arrays have a physical layout where each electrode’s position affects how it interacts with the neurons. Images map onto that grid more readily than text or language, which made visual information a natural starting point.

A business use case emerged from there. By modeling how neurons process images, TBC believes the insights can improve AI models for generating video.

Jeff Dean’s Suggestion

Pomeraniec says Jeff Dean, a prominent AI researcher and TBC investor, was the person who first suggested the company try fine-tuning video generation models before applying its neural tech more broadly. Dean pointed to existing industry benchmarks for video models, which let TBC test early performance against problems that had already been solved.

“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 explains.

What the Startup Claims

Previously, TBC’s technology was only available through a neocloud provider called Bluesky Compute. The startup says its model is up to five times faster at video generation and cuts the cost of inference, the processing part of an AI model, compared to training. It will not say which open-source AI it compares against, but calls it one of the “frontier” video generation models.

Amazon’s Track Record With Wetware

Amazon already offers a biologically derived computing platform from Cortical Labs, an Australia-based company that combines lab-grown neurons with silicon chips. Cortical Labs sells a multi-thousand dollar “biological computer” designed for labs that can reportedly keep neurons alive for six months.

Deap Ubhi, global director of technology for startups at Amazon Web Services, says TBC took a “pragmatic approach” that appealed to Amazon. Instead of swinging for the transformer moon, TBC is working within existing standards of the generative AI space and asking how to make current visual models more efficient.

Scaling the Living Computers

Ubhi says he is excited about TBC but notes the question of scale remains open. He wonders whether pushing the tech to extremes — longer videos, ten minutes or an hour — would hold up in fidelity over time.

“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?” he asks.

What Comes Next

TBC’s tech is now distributed through AWS, which puts it in front of a much larger customer base than Bluesky Compute ever reached. The company’s path so far follows a clear order:

  1. Raise $25 million from Primary Venture Partners, its first significant round of funding.
  2. Secure an additional $25 million shortly after, bringing its total to more than $50 million.
  3. Open offices and a small research and development lab in San Francisco.
  4. Use multi-electrode silicon arrays made by 3Brain, a Swiss biotech company.
  5. Test its tech against existing video benchmarks.

Whether the rat neurons can scale to enterprise video work is the question everyone is waiting to answer.

Source material: “AI Models Built From Rat Brains Just Got Closer to Reality,” WIRED.

The Notebook

Get the Notebook.

The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.

We send one note to confirm. Every issue has a one-click way out.

Advertisement

Leave a Reply

Your email address will not be published. Required fields are marked *

As an Amazon Associate, Clay Tribune earns from qualifying purchases.