Manifesto

A research institution with a public surface.

Phi9 is building physical intelligence through research, data systems, and public method rather than generic AI company language.

Phi9 is a research institution with a public surface, not a pitch deck.

We are building physical intelligence through data systems, method, and deployable work. That means the institution has to be legible from the outside. It has to show what it believes, what it is building, and what kinds of questions it is serious enough to stay with.

Presence

We do not want the language of generic AI companies.

We want language that can hold work, method, evidence, and responsibility. The public surface should feel authored, precise, and calm. It should read like a place where real things are being built.

Method

We believe physical AI advances through loops, not isolated artifacts.

Capture, structure, training, evaluation, and deployment are not separate departments pretending to be progress. They are one system. If the loop is broken, the model story is usually broken too.

Data

We believe data should carry intent, not just observation.

A useful physical dataset is not a folder of recordings. It is a structured representation of action, task, context, failure, and reuse. That is why data is not secondary work here. It is part of the research core.

Proof

We prefer artifacts to claims.

If a page cannot show what the work is, how it operates, or why it matters, then the page is probably saying too much. Research should publish traces. Products should show their operating logic. Infrastructure should feel concrete.

Open work

Some of the work will stay unfinished in public.

That is acceptable. Open questions, partial systems, failed assumptions, and evolving methods are part of the institution. They belong on the surface when they help make the work more legible.

Commitment

The goal is not spectacle. The goal is transfer.

A benchmark matters only if it predicts behavior outside itself. A demonstration matters only if it can be reused. A system matters only if it reduces the distance between research and operation.

We are working toward systems that can operate reliably for real workflows, with better data, better evaluation, and better transfer from research to deployment. The institution should make that ambition clear without inflating what has already been solved.

Contact

Working toward physical intelligence?

If this aligns with what you are building — in data, research, or deployment — write to us.

Read the research