NeoCognition Raises $40 Million to Develop Self-Learning AI Agents

NeoCognition has emerged from stealth with $40 million in seed funding to build self-learning AI agents that can specialize in specific domains, announced in a press release. The seed round was co-led by Cambium Capital and Walden Catalyst Ventures, with participation from Vista Equity Partners and several angel investors, including Intel CEO Lip-Bu Tan and Databricks co-founder Ion Stoica.

NeoCognition was founded by Yu Su, Xiang Deng, and Yu Gu. Su, who leads an AI agent research lab at Ohio State University, serves as the company’s CEO. The team has previously developed research projects such as Mind2Web, MMMU, and SeeAct, which have influenced large language model development by companies like OpenAI, Anthropic, and Google.

The company’s agents are designed to continuously learn the structure, workflows, and constraints of their operating environments. This allows them to become domain experts by building a model of their specific micro-worlds. NeoCognition’s approach is intended to produce agents that are faster, more cost-effective, and more reliable than current general-purpose systems.

According to the announcement, NeoCognition’s technology aims to enable enterprises to deploy agents capable of mastering specialized tasks with minimal manual customization. The company is headquartered in Palo Alto, California, and currently employs about 15 researchers.