NeoCognition

AI agent research lab developing self-learning agents that specialize into domain experts. Rather than relying on static, general-purpose agents, NeoCognition's systems build a "world model" of a specific work environment — its structure, workflows, and constraints — and keep learning on the job to become faster, more reliable, and more cost-effective over time. The company sells its agent systems to enterprises and established SaaS vendors, who can use them to deploy AI workers or embed specialized automation into existing products. Spun out of an Ohio State University AI agent lab in 2025 and led by co-founder and CEO Yu Su, a Sloan Research Fellow, alongside co-founders Xiang Deng and Yu Gu; the roughly 15-person team is largely PhD researchers. No product is publicly available yet, and the technical approach has not been disclosed in detail beyond the world-model framing. Emerged from stealth in April 2026 with a $40M seed round co-led by Cambium Capital and Walden Catalyst Ventures, with Vista Equity Partners participating and angel backers including Intel CEO Lip-Bu Tan and Databricks co-founder Ion Stoica.