Key facts
- The Biological Computing Co. (TBC) uses living neurons to improve AI models.
- TBC and AWS are collaborating to bring neuron-derived AI video models to market.
- TBC's optimized text-to-video model offers 80% lower inference costs and 5x faster generation.
- The model is enhanced by a lightweight software layer derived from living neural activity measurements.
- The software layer adds less than 0.1% to the underlying model and runs on conventional AI infrastructure.
- TBC plans to run its optimized model on AWS Trainium and deploy it through Amazon SageMaker AI.
The Biological Computing Co. (TBC), a San Francisco-based startup, is collaborating with Amazon Web Services (AWS) to integrate its neuron-derived AI optimization technology into cloud-based AI models. The partnership aims to deliver AI video models that are five times faster and 80% cheaper to run, utilizing insights from living neurons to enhance computational efficiency.
TBC's approach involves using real neurons, some sourced from rat brains and others grown from human skin cells, to discover new algorithms. These algorithms are then translated into a lightweight software layer that enhances conventional AI infrastructure without requiring biological hardware or new customer workflows. This software layer adds less than 0.1% to the underlying AI model.
Alex Ksendzovsky, CEO and co-founder of TBC, stated that the partnership with AWS will scale neuron-derived AI optimization commercially, turning biological discoveries into faster, cheaper AI for businesses and creators. Jason Bennett, Vice President at AWS, highlighted that TBC's approach learns from the human brain to make AI more efficient and economical, and that AWS will support TBC's commercialization through its AI stack and marketplace.
For generative AI deployments, inference efficiency is crucial for product economics, allowing platforms to serve more users at a lower cost per output. Faster generation enables quicker iteration for creators and businesses, while improved quality reduces wasted compute on unusable outputs. TBC plans to run its optimized models on AWS Trainium and make them available through Amazon SageMaker AI and AWS Marketplace.