Halluminate is building a suite of products and services to help AI labs train computer use agents (e.g. realistic environments, proprietary datasets, evaluations).
Halluminate
Halluminate is building the training ground for agentic AI. Their platform provides foundation model labs and enterprises with high-fidelity, resettable environments and curated datasets that allow advanced “computer-use” agents to be trained and evaluated safely, reproducibly, and at scale. Instead of testing agents on flaky, live websites, teams can simulate complex workflows—like deal sourcing in a synthetic CRM—inside deterministic sandboxes.
As AI labs like Anthropic, OpenAI, and Meta push billions of dollars into agent development, the need for reproducible environments is no longer optional—it is existential. Halluminate is positioning itself as the indispensable picks-and-shovels provider for the agentic AI wave. We invested because of their sharp insight that use cases, not raw data, are the real bottleneck. Cofounders Wyatt and Jerry met while studying computer science at Cornell and have been living and working together for over seven years. They’re exceptional builders plus authentic people you just trust from day one.
Founders
Previously, he led product and research at Capital One Labs, launching one of the first AI agents in financial services and co-authoring three patents. Jerry studied Computer Science and Economics at Cornell
Wyatt also studied computer science and gained multiple experiences as a software engineer at data infrastructure companies.