OpenAI is backing a new startup called Isara, which is building software that allows thousands of AI agents to communicate and solve complex problems together.
According to a report by The Wall Street Journal, Isara has raised $94 million from investors including Amity Ventures, Michael Ovitz, and Stanley Druckenmiller. OpenAI backed the company at a $650 million valuation.
The San Francisco-based startup was founded in June 2025 by two 23-year-old researchers: Eddie Zhang, a former OpenAI AI safety researcher who previously held a visiting research position at MIT, and Henry Gasztowtt, a computer science student at Oxford University.
The duo co-authored a paper at ICML 2024, before founding Isara, exploring how AI systems could cooperate to improve policymaking — the intellectual foundation on which the company is built.
Since its founding, the company has hired approximately twelve additional researchers from Google, Meta, and OpenAI.
What does Isara actually do?
Isara’s core focus is building the underlying architecture that allows different AI agents to align goals, exchange data, and divide tasks across complex industrial processes — functioning, as the founders describe it, like a well-coordinated team rather than isolated tools.
Unlike traditional AI models, such as LangChain, CrewAI, and AutoGen, which operate in isolation, Isara’s vision is to create swarms of agents that collaborate to tackle large-scale problems no single model can handle alone.
The startup is initially targeting investment firms with predictive modelling software, with biotechnology and geopolitical analysis as secondary markets. The longer-term goal, as reported by WSJ, is to train these agents to collaborate on challenges such as tracking geopolitical shifts or forecasting economic trends.
Isara is part of a growing wave of so-called “neolab” AI research startups attracting significant investor attention. According to The Information, investors have poured or discussed $2.5 billion across five such startups in just over a month, reflecting a broader race to back foundational AI research outside of the major labs.
