Anomity Joins the NVIDIA Inception Program

In September 2026, Anomity's application to the NVIDIA Inception program was approved.

As with our earlier program posts, we want to be clear about what this is. A startup program is not a security certification, not an audit and not a statement by NVIDIA about our product. It is access, resources and community. One part of it does bear on what we build, and that part is worth explaining.

What the program includes

• Technical resources and training, including personalized recommendations for tools and courses through the member portal.
• Preferred pricing on NVIDIA hardware and software.
• Offers from NVIDIA's partners.
• Opportunities to connect with investors.
• The NVIDIA Developer Program, which complements membership with free access to SDKs, APIs, developer tools, training and technical documentation.

NVIDIA describes the program as free, with no application fees, membership fees or equity requirements. It is not an investment, and the investor network is an introduction channel rather than capital.

Why it matters for coverage

Anomity's core claim is a complete inventory of the AI artifacts running across a fleet: agents, MCP servers, extensions, plugins, skills, secrets, hooks, CLIs and local LLM runtimes. That claim has a shelf life. Every time a vendor ships a new kind of runtime or agent surface, the inventory is incomplete until we recognize it. The gap between a surface shipping and a discovery tool seeing it is coverage lead time, and it is dead space during which security teams believe their picture is complete when it is not.

A growing share of that surface runs on NVIDIA. Local LLM runtimes run on GPU workstations and internal servers, and they are the category where hosted-model safeguards are absent by construction, as Anthropic's GLM-5.3 study made clear. And in the same month our membership was approved, NVIDIA shipped OpenShell and the Open Agent Safety Platform, a runtime that sits between agents and everything they reach. Each of these is something to discover before it can be governed. Access to NVIDIA's developer resources shortens our lead time on them, and we intend to spend it on coverage.

What it is not

• Not an endorsement. NVIDIA has not reviewed or certified Anomity's product.
• Not an integration. Anomity does not integrate with OpenShell or Sentry today, and membership does not change that by itself.
• Not a change to data handling. Anomity remains SOC 2 Type II attested, collects metadata only, and redacts secret values on the endpoint before anything leaves the machine.

Where it fits with our other programs

The NVIDIA Inception program joins two Anthropic programs. Anomity was approved in Anthropic's Cyber Verification Program in June 2026, which is the security-relevant one: an application-based program for dual-use defensive work. In August we joined Claude for Startups, which, like this program, is access and community rather than capital. Anomity is also a signatory of OpenAI's open letter, A call for collective action on cyber defense.

The common thread is the same each time: early sight of the surfaces our customers will have to govern, so the inventory is right on the day those surfaces reach their fleets rather than months later.

How Anomity uses this lead time

The Endpoint Sensor inventories the AI artifacts on every managed endpoint, including the local LLM runtimes and agent CLIs that run on GPU workstations. The Browser Sensor covers 311 AI web services and whether the account in use is corporate or personal, and cloud discovery covers the Google Workspace and GitHub OAuth grants held by AI applications. On agents that expose a hook, such as Claude Code's PreToolUse, Anomity decides allow, deny or log before a tool call runs, and every decision lands in a 90-day audit trail.

Early sight of new surfaces is only useful if it turns into coverage. That is what we intend to do with it. To see what is running across your own fleet, book a 30-minute demo.