AI Models

This AI Video Model Learned From Real Neurons—and Cuts Inference Costs by 80%

Importantly for enterprises, the approach does not require biological hardware.

(Image-Magnific)

The Biological Computing Co. (TBC), a San Francisco-based company using living neurons to study more efficient ways of computing, has partnered with Amazon Web Services (AWS) to bring what it describes as the world’s first neuron-derived AI video model to market.

The company said its optimised text-to-video model delivers 80% lower inference costs and five-times faster generation than the underlying open-source video model, while also improving output quality. The optimisation comes from a proprietary software layer derived from measurements of neural activity in living neurons.

Importantly for enterprises, the approach does not require biological hardware. TBC said the software layer adds less than 0.1% to the underlying model and runs on conventional AI infrastructure.

“Our partnership with AWS takes neuron-derived AI optimization to commercial scale. We’re turning discoveries from real neurons into faster, cheaper AI for businesses and creators. More importantly, biology gives us a fundamentally different engine for discovering better optimisation strategies over time,” said Alex Ksendzovsky, TBC Co-Founder and CEO.

Under the partnership, TBC plans to run its optimised model on AWS Trainium, make it available through Amazon SageMaker AI and pursue commercial distribution through AWS Marketplace. This could allow customers to access the optimised models through existing AWS infrastructure and workflows.

I met with Alex at their lab yesterday and got a ‘hands-on’ tour of the innovation they are working on. 😷 Cool stuff. 🤓 https://t.co/KqxgIUZ8jE pic.twitter.com/NSMdI4U059
— Thomas Dohmke (@ashtom) September 23, 2026

“Nature solved the computing efficiency problem billions of years ago. TBC's insight is that we can learn from the original computer—the human brain—to make AI faster, more efficient, and more economical. By building on the AWS AI stack and leveraging go-to-market channels like AWS Marketplace, TBC can move from discovery to commercial scale at startup speed,” said Jason Bennett, AWS VP & Global Head, Startups and Venture Capital.

The announcement also comes as AI infrastructure companies and model developers look for ways to improve efficiency rather than relying solely on larger models and additional compute.

“Compute is becoming one of the biggest constraints on AI. We need more infrastructure, but we also need to make every unit of compute dramatically more productive,” said Jon Pomeraniec, TBC Co-Founder and COO.

TBC said the video model is the first commercial product from its biological-computing research and plans to apply insights from its neural experiments to other AI architectures and workloads.