Subquadratic

Category
Frontier AI Model

Status
Active

Leadership
James McGoff - Founder & CEO

Subquadratic builds frontier AI models that reason across millions of tokens at a fraction of the compute. Their architecture, Subquadratic Sparse Attention, reads only the relationships that matter, so one model can hold an entire codebase, a decade of filings, or a full patient history in a single pass.

Subquadratic came out of an early Pave position, and its team went and solved the efficiency problem every lab is spending billions to chase. The layer they built is the one real consumer health AI will run on. Multi-million token reasoning across records, research, and a lifetime of personal data is what turns AI-guided care from a demo into something affordable at consumer scale. We moved at the ground floor, and we're on the cap table of one of the most-watched models in the market.