The Series A will accelerate Raindrop’s anomaly detection research, scale the product to more enterprises, and advance Simulations, a new product that tests agent changes against production traffic before they ship

Raindrop, the agent reliability company, raised a Series A led by CRV, bringing total funding to $50 million. Existing investors Lightspeed Venture Partners, Y Combinator participated along with lead researchers from OpenAI, Anthropic, and Thinking Machines. Raindrop is trusted by some of the largest and fastest-growing companies in the world, including Vercel, Framer, Clay, and Fortune 100 Enterprises to protect their users from agent failures.

Today, Raindrop is also announcing Simulations, currently in research preview. Traditional evals depend on pre-defined test cases and primarily measure failures a team has anticipated. Simulations replay real production traffic and existing test cases against a proposed change to an agent harness, then apply Raindrop’s anomaly detection to the results. Teams can measure performance against their own test cases while also detecting unexpected behavior changes before they reach production. In other words, AI engineers can now see “what their change will change.”

Agents are becoming more capable and complex. The research group METR found that the length of tasks agents can complete on their own doubles roughly every seven months, and a single run can now last days and involve thousands of tool calls.

“Agents now run for hours, call thousands of tools, and handle real money, real health data, and real customers,” said Zubin Koticha, CEO of Raindrop. “When an agent fails, it does the wrong thing convincingly at scale until someone happens to notice. Raindrop catches those failures in production, and with Simulations, before a change ever ships. This funding lets us bring the testing process frontier labs use on their own models to every team building agents.”

Raindrop detects semantic anomalies in production agent traffic. Issues show up in traces before a user ever complains. When behavior shifts, engineering teams see what changed, when it started, and which users it affected, with hundreds of real examples attached.

Simulations mark the first time companies have access to the same training and testing process used by frontier labs. OpenAI recently published research on deployment simulation, which regenerates responses to de-identified production conversations with a candidate model to predict misbehavior rates before release. Anthropic builds synthetic universes to train and stress-test its agents.

“I’ve spent a decade backing critical infrastructure for developers,” said Reid Christian, General Partner at CRV. “Agents are fundamentally different from traditional software. They are highly capable, autonomous, and non-deterministic. Raindrop treats agent failure as a detection problem, the way a security company would, and Simulations close the loop from production to development. We’re thrilled to lead this round.”

“As agents are deployed across both everyday products and into high-stakes environments like defense, ‘bad behavior’ will increasingly become catastrophic,” said Bucky Moore, Partner at Lightspeed. “We backed this team at the seed because they demonstrated incredible product vision and instincts, and translated this into a level of customer delight we rarely see at such an early stage. A year later, Fortune 100 companies are running Raindrop on their agent traffic. Lightspeed is proudly doubling down on Raindrop as the leader in this new category.”

Raindrop’s team combines world-class ML and product talent, including engineers who invented fraud-transformer models at Robinhood and pioneered malicious anomaly detection at Square, senior security engineers from Segment, Semgrep, and Socket.dev, engineers who have led teams across Apple, Airtable, and DoorDash, and designers from Apple and PlanetScale. Raindrop was founded by Zubin Koticha, Ben Hylak, and Alexis Gauba.

“If we’re having an issue like a build failure or agents stuck in a loop, we see that issue in Slack. We see context like how many people are experiencing it and chats we can dig into. More than anything, Raindrop brought more visibility into the most severe issues,” said Bani Singh, AI Engineer, Vercel.