Raindrop, the agent reliability company, has announced a Series A funding round led by CRV that brings its total funding to $50 million. The investment included participation from existing backers Lightspeed Venture Partners and Y Combinator, along with lead researchers from OpenAI, Anthropic, and Thinking Machines. The company 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.
Funding and Strategic Support
Reid Christian, General Partner at CRV, stated that he has spent a decade backing critical infrastructure for developers. He noted that agents are highly capable, autonomous, and non-deterministic, which makes them fundamentally different from traditional software. Christian added that Raindrop treats agent failure as a detection problem, similar to a security company, and that Simulations close the loop from production to development.
Introducing Simulations
Alongside the funding, Raindrop introduced Simulations, a product currently in research preview. Traditional evaluations rely on predefined test cases and mostly catch failures that teams anticipated. Simulations replay real production traffic and existing test cases against proposed changes to an agent harness, then apply Raindrop's anomaly detection to reveal unexpected behavior changes before deployment.
Growing Complexity of AI Agents
The need for such tools is growing as agents become more capable and complex. Research group METR found that the length of tasks agents can complete on their own doubles roughly every seven months, with single runs lasting days and involving thousands of tool calls. Raindrop detects semantic anomalies in production agent traffic and shows engineering teams what changed, when it started, and which users were affected.
Leadership Perspective
Zubin Koticha, CEO of Raindrop, said agents now run for hours, call thousands of tools, and handle real money, real health data, and real customers. He explained that when an agent fails, it can do the wrong thing convincingly at scale until someone notices. Koticha added that the funding allows Raindrop to bring the testing process used by frontier labs to every team building agents.
Customer and Investor Validation
Bani Singh, AI Engineer at Vercel, said Raindrop provides visibility into severe issues such as build failures or agents stuck in a loop, with context on how many people are affected. Bucky Moore, Partner at Lightspeed, noted that Fortune 100 companies are now running Raindrop on their agent traffic. Lightspeed doubled down on its investment because of the team's product vision and rare early-stage customer delight.
Team and Market Position
Raindrop's team includes engineers who invented fraud-transformer models at Robinhood and pioneered malicious anomaly detection at Square, alongside leaders from Apple, Airtable, and DoorDash. The company was founded by Zubin Koticha, Ben Hylak, and Alexis Gauba. Simulations are now available in research preview, and teams can request access through the company's website.
With its Series A and $50 million in total funding, Raindrop is positioned to expand anomaly detection research and scale its platform to more enterprises. The company is also advancing Simulations to help teams test agent changes against production traffic before they ship. As AI agents take on longer and higher-stakes tasks, Raindrop's focus on detecting silent failures addresses a growing reliability challenge for organizations worldwide.