San Francisco – September 22, 2026 -- The Biological Computing Co. (TBC) has partnered with Amazon Web Services (AWS) to bring what it calls the world's first neuron-derived AI video model to market, cutting inference costs by 80% and speeding generation by 5x compared to the base open-source model.
TBC's software layer runs on standard AI infrastructure without new hardware
The optimization is delivered through a proprietary software layer derived from measurements of living neural activity. The layer adds less than 0.1% to the underlying model's footprint, requires no biological hardware, and demands no change to existing customer workflows, according to TBC.
AWS commits Trainium, SageMaker AI and Marketplace to scale the technology
Under the collaboration, TBC plans to run its optimized model on AWS Trainium, make it deployable via Amazon SageMaker AI, and distribute it commercially through AWS Marketplace. The arrangement lets customers access TBC-optimized models inside AWS environments they already use.
"Our partnership with AWS takes neuron-derived AI optimization to commercial scale," said Alex Ksendzovsky, CEO and co-founder of TBC. He added that biology provides "a fundamentally different engine for discovering better optimization strategies over time."
AWS frames the deal as a bet on biology-inspired computing efficiency
Jason Bennett, Vice President and Global Head of Startups and Venture Capital at AWS, said TBC's approach of learning from "the original computer -- the human brain" allows the startup to move from discovery to commercial scale at startup speed using AWS's go-to-market channels.
Lower inference costs address a widening compute constraint for AI deployment
Jon Pomeraniec, co-founder and COO of TBC, said compute is becoming one of the biggest constraints on AI, and that lower inference costs let more companies afford to build and scale AI systems. For platforms running generative video at volume, the 80% cost reduction and 5x speed gain directly affect unit economics, allowing more users to be served on the same infrastructure while cutting failed or unusable generations.
Video model is TBC's first commercial product from an ongoing discovery pipeline
This marks the first commercial product from TBC's discovery process, which builds a growing library of neural-response data and potential algorithms from biological experiments. TBC said it plans to apply the same process to additional AI models and architectures, with further applications across other AI workloads to follow.