What it sells
The cheapest, fastest inference cloud
Business model breakdown
Buyer: who owns the budget?
UnverifiedML/infra lead or CTO at teams running long-context LLM inference.
Trigger: what makes them buy now?
High prefill cost and slow time to first token on long-context requests.
Workflow: what inputs become what output?
Prefill runs on a small model and KV cache is transferred to a larger model for decode; or use serviced private open-weight models with a one-line config change.
Economics: what it saves and how it charges
Claims 20% lower inference cost and time; benchmark 84% cheaper prefill ($46.9k to $7.5k per 1M requests) and 6.3x faster first token on 100k-token context.
Distribution: how it reaches customers
Demo-led (Book a demo), managed serving of open-weight models.
Funding history and coverage
| Round date | Stage | Amount | Investors | Source |
|---|---|---|---|---|
| — | Y Combinator Summer 2026 | — | Y Combinator | ycombinator.com |
Profiles summarize public reporting and company websites. Business-model notes are MaxBC's reading of that material, not confirmed facts; unknowns are left blank rather than guessed. Updated 2026-10-12.