Dipole ML Announces Upcoming October 1 Launch of Sovereign Frontier Intelligence

Dipole ML Inc. will launch publicly on October 1, 2026. Founder Saurabh Sharma (@saurabhsnrg) confirmed the date today, marking the transition from four years of private research to a public frontier AI lab with its own models, products, and infrastructure.

Four years, one savings account

Dipole ML began in 2022 as an intense, independent research effort, funded entirely out of personal savings. There is no venture capital on the cap table, no outside board, and no advertising business model shaping the roadmap. That independence was a deliberate choice: it is the only way to build frontier intelligence whose first obligation is to the people using it.

What launches on October 1

The launch brings three surfaces to the public, all served from Canada:

• The Raven model family: Raven Flash for velocity and Raven Max for deep reasoning, both with text and image understanding, native tool use, and streaming reasoning
• DCode, our agentic coding environment for working inside real codebases
• The Dipole ML API, pay-as-you-go and compatible with existing SDKs

An independent frontier lab

The mission is specific: create an independent frontier AI lab that couples top-tier global performance with statutory privacy. Performance you can verify on public benchmarks, privacy enforced by architecture rather than promise, and sovereignty baked into where the compute lives. Full details arrive with the launch on October 1.