November 6, 2026Funding-Series CAgentsInfrastructure

Sentra Raises $150M Series C to control what AI agents can reach in company data

Capital One Ventures led a $150M Series C in Sentra and takes a board seat. Bessemer, Zeev, Standard Investments, Munich Re Ventures and Key1 Capital are in too. Valuation wasn't disclosed, and total funding now tops $260M. When a bank's venture arm leads a round like this, you can guess what its parent's security team worries about.

Sentra's pitch is that every agent a company switches on is another identity reading its data, often around the clock, inheriting access to data nobody ever classified. So it works on data at rest: classify it, map which people, service accounts and agents can reach what, including inherited permissions, delete the stale stuff, cut access down to the minimum. If an agent reaches for sensitive data it normally doesn't touch, it gets flagged, and permissions can be stripped automatically. After an incident, you get a record of everything that agent could have reached. Analysis runs inside the customer's environment. The growth claim of over 300% is the company's, and the source here is SiliconANGLE's write-up.

The other money this week follows the same logic, one step later. Pliant, the Berlin card and payments company past $100M in recurring revenue per TNW, raised an $80M Series C led by Seaya Nova and Speedinvest to open its B2B payments to Claude, ChatGPT and in-house agents. No valuation seen. Its lead investor's line is the sharpest of the week: "When agents move money, someone still has to be licensed, auditable and liable." Shackleton AI raised a $5M seed led by Battery Ventures for agents in construction. And Nooks took $500M from General Catalyst's Customer Value Fund, which is growth capital against a share of returns, not an equity round at a price, so don't read that one as a valuation event.

The controls are moving into products too. On Nov 5 Anthropic made prompt and pre-tool-call hooks generally available for Claude Enterprise, so existing security tools can check tool calls before they run, plus a security center and an Admin API sign-out in beta. Read the fine print: sign-out doesn't end API keys, cloud sessions already running or scheduled routines. Kill switches keep having that kind of footnote.

And for the data side, NVIDIA's Auto Ontology is GitHub's top trending agent repo today, at about 2,700 new stars with a version bump to 1.2.0. It builds a semantic layer from enterprise metadata and serves it to any harness over MCP, CLI and API. The README's framing is the right one: a revenue metric answered faster is worth nothing if the number is wrong. Reliability claims there are NVIDIA's own. Put it together and the agent data layer now has two questions with real money behind them: what may the agent touch, and does it know what it's looking at. Sources: https://siliconangle.com/2026/11/04/sentra-raises-150m-to-control-what-ai-agents-can-reach-in-corporate-data/ https://sentra.io/ https://thenextweb.com/news/pliant-80m-series-c-ai-agents-b2b-payments-seaya-speedinvest https://www.prnewswire.com/news-releases/shackleton-ai-unveils-agentic-operating-system-for-construction-raises-5-million-led-by-battery-ventures-302928375.html https://www.nooks.ai/blog-posts/introducing-node-revenue-agent-platform https://claude.com/resources/articles/claude-enterprise-security-updates https://github.com/NVIDIA/auto-ontology
← Previous
Uber's AI agent numbers from one of its biggest weeks: 320 million trips and orders, 4 rides
← Back to all articles

Comments

Loading...
>_