The Biological Computing Co. (TBC) is a US company using living neural networks to discover computational principles that can be translated into conventional AI software. It emerged from stealth in February 2026 with a $25 million seed round led by Primary Venture Partners.
Its current commercial route is software-first. Researchers encode model-relevant problems into living neural networks, record the resulting dynamics and derive lightweight software adapters from useful patterns. The customer can then run the adapted model without maintaining living cells.
AWS collaboration announced in September 2026
In July 2026 TBC released an interactive Neural Optimizer proof of concept for the Oasis 500M video model. On 23 September it announced a collaboration with Amazon Web Services to take a neuron-derived text-to-video model towards commercial deployment. TBC now reports 5× faster generation and 80% lower inference cost against its chosen base model, with improved output quality. Those figures are company measurements and have not been independently benchmarked.
The deployable layer is small conventional software: TBC says it adds less than 0.1% to the underlying model and runs on ordinary AI infrastructure. The collaboration is commercially relevant because customers do not need living-neuron hardware in their own inference stack.
Where biology sits in the stack
Today, the neurons are part of algorithm discovery. TBC says its longer-term direction is real-time biological compute operating alongside silicon. That distinction matters: the current product can be deployed as software even though living neurons were involved in discovering the optimisation.
What would strengthen the case
Independent replication, public ablation studies, fixed benchmark conditions and a detailed account of which biological dynamics transfer into the adapter would make it easier to judge how general the result is.