As AI tools like ChatGPT become widely integrated into daily operations, companies are facing a new kind of security threat, one that comes from in-house. Employees, often without realizing it, are exposing sensitive data by feeding proprietary information into AI systems. San Jose-based Cyberhaven, now valued at $1 billion, is stepping in to solve that problem.
Fresh off a $100 million funding round, Cyberhaven uses AI to watch how data moves through a company in real time. Its tools, Linea AI and a proprietary Large Lineage Model (LLiM) have analyzed hundreds of billions of dataflows across client networks. This data history allows the platform to understand what “normal” looks like and detect when something seems off, like a user pasting sensitive product specs into a generative AI model.
According to CEO Howard Ting, the system predicts what a user is likely to do next with a given piece of data. If someone suddenly does something out of character, like posting confidential documents to a personal cloud, the software flags it. And instead of overwhelming security teams with alerts, Cyberhaven prioritizes threats based on how serious the data exposure could be.
Just a year ago, the company was valued at $400 million after raising $88 million. Now, with participation from StepStone Group, Schroders, Industry Ventures, Khosla Ventures, and others, its valuation has soared. And unlike many startups that quickly sell to bigger players, Cyberhaven plans to stay independent, confident in its unique technology and market opportunity.
Big names are already using Cyberhaven, including Motorola, Oscar Health, and Cooley. The company expects to generate over $50 million in recurring revenue this year. Khosla Ventures’ Ethan Choi says the platform protects what matters most: customer data, internal blueprints, even design files.
“You’re finally able to trace and protect the Crown Jewels wherever they are—in rest, in movement, or even inside AI models,” he explains.
Founded in 2016 by five PhDs and initially funded by DARPA, Cyberhaven began as a tool for securing intellectual property. Over time, it evolved into a broader AI-powered security platform. What sets it apart is its precision. Unlike traditional tools that flag everything, Cyberhaven zeroes in on what’s actually important, cutting down on false positives while increasing protection.
That focus on control over chaos has real-world implications. Oscar Health, for example, uses Cyberhaven not to block AI usage, but to guide employees. Through browser extensions and a desktop app, employees are told what they can and can’t share with AI systems. Things like personal medical data? Off-limits. But the company encourages AI exploration, within safe boundaries.
Cyberhaven’s real power might go even further than data security. By mapping how data moves within an organization, the platform could also surface insights about team productivity and workflow bottlenecks. CEO Ting suggests it could even reveal why one engineering team outperforms another, insights that can lead to better business decisions.
At a time when data is not just valuable but vulnerable, Cyberhaven is helping companies understand where their data goes, how it’s used, and how to protect it in an AI-driven future.
