Raindrop Raises Series A to Catch AI Agents Failing

Raindrop said Sept. 17 it has raised a Series A round led by CRV, taking its total funding to $50 million. The company's own announcement does not break out how much of that total came from the new round versus its earlier seed funding, and TheNextWeb, which also covered the raise, did not report a separate figure either.

Raindrop monitors AI agents while they run in production, watching for what TheNextWeb described as semantic anomalies: hallucinated answers, an agent calling the wrong tool, or its behavior shifting after the underlying model gets swapped out. Chief Executive Zubin Koticha, one of the company's three co-founders along with Ben Hylak and Alexis Gauba, told TheNextWeb that "agents now run for hours, call thousands of tools, and handle real money, real health data, and real customers," and that "when an agent fails, it does the wrong thing convincingly at scale until someone happens to notice." Raindrop's own site says it processes billions of traces a month for Fortune 100 customers.

What Raindrop's monitoring flags	Example
Hallucinated answers	An agent states something its retrieved data does not support
Tool misuse	An agent calls the wrong tool, or the right tool with the wrong inputs
Behavior drift	An agent's output pattern changes after a model swap it wasn't tested against

The product Raindrop launched alongside the round

Raindrop used the funding announcement to open early access to Simulations, which replays real production traffic and existing test cases against a proposed change to an agent's setup, then runs Raindrop's own anomaly detection against the results, according to the company's post. The pitch is that a team can see whether a change altered agent behavior before merging it rather than after a customer notices. Named customers cited by TheNextWeb include Vercel, Framer and Clay, plus unnamed Fortune 100 companies in healthcare and logistics; Vercel AI engineer Bani Singh told TheNextWeb, "If we're having an issue like a build failure or agents stuck in a loop, we see that issue in Slack."

Who else is behind the round

Alongside CRV, Raindrop's announcement names Lightspeed Venture Partners and Y Combinator as returning investors, plus individual researchers and executives from OpenAI, Anthropic and Thinking Machines investing personally; TheNextWeb also reported Figma Ventures and Vercel Ventures among the backers. CRV general partner Reid Christian told TheNextWeb that "Raindrop treats agent failure as a detection problem, the way a security company would," a framing that puts Raindrop in the same category as the vendors that scan code for vulnerabilities rather than the observability tools that only chart uptime. Raindrop's raise lands inside the same AI-agent boom that has drawn scrutiny for security flaws across agent skill marketplaces and for OpenClaw's own security debt as adoption outran review.

What Raindrop hasn't said

Neither Raindrop's announcement nor TheNextWeb's coverage discloses a valuation, the size of Raindrop's seed round, or the company's revenue. The next disclosure to watch is whether Raindrop reports usage numbers for Simulations once it leaves early access, the kind of figure that would show whether teams are actually running proposed agent changes through it before shipping, rather than just after something breaks.

Sources

1. Announcing our Series A and $50M in total funding to protect the world from AI agent failures, Raindrop
2. Raindrop Series A takes it to $50m for agent monitoring, TheNextWeb