NeoCognition has emerged from stealth with a $40 million seed round as it pushes to build AI agents that learn and specialize through experience. The Palo Alto startup was founded by Ohio State University professor Yu Su alongside Xiang Deng and Yu Gu.
Cambium Capital and Walden Catalyst Ventures co-led the round, while Vista Equity Partners joined alongside angel backers including Intel CEO Lip-Bu Tan and Databricks co-founder Ion Stoica.
NeoCognition aims to address one of the biggest weaknesses in AI agents, which is reliability. Su has argued that many current agents still fail to complete tasks consistently, which limits how far companies can trust them in real work settings.
He founded the startup after concluding that recent advances in foundation models made more personalized agents possible.
The company announced the financing on April 21 and said the round was oversubscribed. In addition to its lead investors, the cap table includes enterprise-focused investor Vista Equity Partners and several AI researchers and operators. NeoCognition said the funding will support work on agents that learn continuously and reach expert-level performance over time.
NeoCognition’s product strategy centers on agents that build a ‘world model’ of the environment where they operate. The company says these systems learn workflows, constraints, and relationships inside a specific setting, then use that knowledge to improve performance in the same way workers gain expertise on the job.
NeoCognition positions that approach as an alternative to broad but inconsistent general-purpose agents. The startup plans to sell mainly to enterprises and established SaaS vendors. NeoCognition says customers can use its systems to create AI workers or strengthen existing software products with more specialized automation.
The company also said a deeper understanding of a work environment could improve speed, cost efficiency, reliability, and safety in higher-stakes use cases.
NeoCognition ties its commercial story to Su’s academic record in AI agents. The company said Su’s Ohio State lab produced work such as Mind2Web, MMMU, and SeeAct, which it describes as part of the foundation of the modern agent field. Walden Catalyst’s Lip-Bu Tan said Su’s team has worked across major parts of agent development, including perception, memory, planning, evaluation, and safety.
Su serves as an associate professor and innovation scholar at Ohio State, where he works on language agents, planning, world models, memory, and evaluation. His profile states that he co-directs the OSU NLP group and leads research teams tied to foundational AI and machine learning. Ohio State has also noted that Su joined the university in 2020 and previously worked at Microsoft Semantic Machines.
This funding round reflects strong investor appetite for startups that focus on the reliability layer of AI rather than only on larger foundation models. NeoCognition now enters that race with fresh capital, academic credibility, and an enterprise pitch built around specialized AI agents that keep learning after deployment.
Also Read: Startup News Today: Krafton, Naver Join Mirae Asset for $600M India-Focused Investment Fund