About Neuraphic
Neuraphic is an artificial intelligence laboratory, incorporated as a Delaware C Corporation in the United States. It works on problems that do not have a known solution, and it publishes what it finds. Where a result turns out to be useful to someone outside the laboratory it becomes a product, and that is the only way anything here becomes a product.
The laboratory runs its own infrastructure. It writes its own identity and account systems, operates its own database and holds its own keys, rather than renting those parts from a vendor. That is a research position rather than a security one: a system you did not build is a system whose failures you can only watch, and the failures are the subject.
Journalists are welcome to quote the two paragraphs above verbatim as boilerplate. If you need one sentence: Neuraphic is an artificial intelligence laboratory that studies the limits of what these systems can be trusted to do, and turns what it solves into products.
Press contact
For media inquiries, interview requests and fact-checking, write to press@neuraphic.com. Put your deadline in the subject line and we will work to it. We would rather answer a fact-check in an hour than correct a story after it runs, so if you are checking a claim about us, say so and it goes first.
Anything that is not press goes to hello@neuraphic.com. Product and account problems go to support.neuraphic.com, where they become a tracked ticket instead of an email someone has to remember.
Founding facts
Company name: Neuraphic, Inc. Legal structure: Delaware C Corporation, United States. Incorporated 6 April 2026. Member of NVIDIA Inception since May 2026.
Neuraphic has not raised outside capital and has no investors. That is a fact about the company rather than a stage it is passing through, and we would rather see it reported than left as an implication.
We do not disclose headcount, and we will not confirm or deny third-party reporting on it. Everything else on this page is checkable. If you find something here that is wrong, tell us and we will correct the page rather than argue with your copy.
Recent announcements
The alarms were working. None of them could reach a human. — 11 May 2026
Joining NVIDIA Inception — 5 May 2026
Responsible Scaling Policy, v1.0 — 4 May 2026
Full newsroom
People
Nobody who works here is named publicly, and we will not confirm the name, role or background of anyone, including the founder. This is a standing position rather than a response to any particular story. There are no headshots to send: if a story needs an image, take the mark or the wordmark from the brand assets page.
Quotes are attributed to a Neuraphic spokesperson. We do not offer named executive comment, and an article that needs a person's name attached to a claim about this company will not get one from us.
Brand assets
Logo files, the wordmark, color values, and typography guidance are available on our brand assets page, including a downloadable brand kit. For custom formats or usage questions, contact brand@neuraphic.com.
How we talk about ourselves
A laboratory, not a cybersecurity company. This is the correction we make most often, and for a while this site invited it. Neuraphic does not sell a security product and has none to sell. The work touches security because a system you cannot trust under attack is a system whose limits you have not found, and the limits are the subject.
Safety is the research, not a department. There is no safety team that reviews the work at the end. What a system can be trusted to do is the question being studied, so a story that separates our research from our safety work has divided something that is not divided.
Products are an output, not the plan. There is no product roadmap to describe. Something becomes a product when a result is solid enough to be useful outside the laboratory, and until then it does not exist publicly. That is why there is no early-access list and nothing to brief on under embargo.
Nothing is announced before it runs. If it is in the newsroom, it exists and it is live. We have no unreleased products to preview and we do not comment on what we might build.
The name. Always "Neuraphic", one word, capital N, no space. Never "Neuraphic AI", never "Neura", never an acronym, never set in all caps.