Boltzbit has introduced General Learning Intelligence (GLI), an architecture for AI systems that keep learning after deployment, alongside a new brand, the company said on 30 July 2026.
The London-based research company argues that pre-trained foundation models are costly to build and stay static once deployed, unable to improve through use without retraining.
“Training is the real bottleneck in AI, and we’ve seen what live-learning models are capable of in production through our work with select corporations,” said Dr. Jinli Hu, CTO and co-founder of Boltzbit.
Boltzbit said its technology is already in production, powering applications where live learning improves performance and supports more context-aware AI agents for complex decisions.
Later this year the company plans to release its first product suite, which will bring together tools to create, train and host AI models and to build applications and agents on top of them. Boltzbit was founded in 2020 by former Google and Microsoft AI researchers.