Frontier Computing
Frontier Computing is a biocomputing company that grows large-scale neuronal tissue as a next-generation hardware layer for AI training. Its tissue-culture method overcomes the vascularization limit that previously capped biological neural networks at roughly one million neurons, enabling far larger systems with continual learning and sample efficiency. The company provides the substrate, interfacing, and cloud infrastructure for third parties to train on its systems, with early rollouts demonstrating neurons learning to play Frogger.
- Biotechnology
- Artificial Intelligence
- Hardware
Funding
Raised to date
$500KRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
Founders
Founder details are not available yet.
Product
Problem
Biological neural tissue is already being trained on tasks like games and next-token prediction, but wetware has been limited to roughly one million neurons due to the vascularization problem—past a geometric limit, cells sit too far from an oxygen supply and die. This scale constraint prevents biological systems from competing with or complementing GPU-based training for larger models.
Solution
Frontier has developed a tissue-culture approach that breaks the vascularization limit, enabling the growth of far larger biological neural networks. These systems offer continual learning baked into the substrate, the sample efficiency of living neurons, and a training platform that sits as an abstraction layer above backpropagation. The company provides the substrate, interfacing, and cloud infrastructure for third parties to train on its systems, making biocomputing accessible as a next-generation hardware layer.
Target Audience
Primary customers are research organizations, AI labs, and companies seeking alternative or complementary hardware substrates for neural network training, particularly those interested in biological computing approaches.
Features
- Tissue-culture method that overcomes the vascularization limit, enabling biological neural networks far beyond the ~1M neuron scale
- Continual learning capabilities integrated directly into the biological substrate
- Training platform positioned as an abstraction layer above backpropagation for compatibility with existing ML workflows
- Full-stack offering including substrate, interfacing hardware, and cloud infrastructure for third-party training
- Early rollouts demonstrate neurons learning to play Frogger, showcasing real-world task execution