NeoCognition Raises $40M Seed for Human-Like AI Agents
NeoCognition, an AI research lab founded by Ohio State professor Yu Su, announced $40 million in seed funding after coming out of stealth. Su, who runs an AI agent lab at the university, resisted venture capital pressure to commercialize his research at first. He decided to launch the startup last year once he recognized that progress in foundational models could enable truly personalized agents.
The funding round received co-lead investments from Cambium Capital and Walden Catalyst Ventures. Vista Equity Partners joined, along with angel investors such as Intel CEO Lip-Bu Tan and Databricks co-founder Ion Stoica. This capital supports NeoCognition's work on self-learning AI agents.
Challenges with Today's AI Agents
Su pointed out that current AI agents act as generalists. Users must trust them blindly for each task, he explained to TechCrunch. The main problem stems from inconsistency. Tools like Claude Code, OpenClaw, or Perplexity's computer utilities complete intended tasks only about 50 percent of the time, according to Su.
Because of this unreliability, agents cannot yet serve as trusted, independent workers. NeoCognition plans to address this gap. The company develops a system where agents self-learn to gain expertise in any field, mimicking human learning patterns.
Su noted that human intelligence spans broad areas, but its strength lies in specialization. People quickly grasp the specific rules, connections, and outcomes in new settings or jobs. NeoCognition builds agents to follow this model exactly.
Building World Models for Expertise
Stay ahead of the AI curve
The most important updates, news, and content — delivered weekly.
No spam. Unsubscribe anytime.
Su described human ongoing learning as constructing a world model for any job or setting. Agents need to learn on their own to form models of particular environments, he said. This ability to specialize fast represents the key element missing for reliable independent AI operation.
Training agents for specific tasks works, but it requires custom engineering for each area. NeoCognition stands apart by creating generalist agents that self-learn and adapt to any domain.
The startup plans to offer its agent systems to businesses, especially established software-as-a-service providers. These customers can deploy agent workers or upgrade current products. Su emphasized the value of Vista Equity Partners' investment. As a major private equity player in software, Vista offers connections to many companies seeking AI updates for their offerings.
Team and Focus
NeoCognition employs around 15 people now. Most hold PhDs, reflecting its research-oriented approach.
Yu Su brings deep expertise from his role at Ohio State University, where he leads work on AI agents. Cambium Capital focuses on technology investments, often in hardware and software intersections. Walden Catalyst Ventures supports deep tech ventures. Vista Equity Partners manages billions in software assets, with a track record of scaling enterprise solutions. Lip-Bu Tan, as Intel's leader, has backed numerous AI and chip startups. Ion Stoica helped found Databricks, a leader in data and AI platforms used by enterprises worldwide.
This funding arrives amid heavy investor interest in AI researchers starting companies to improve agent reliability and efficiency. NeoCognition positions itself to fill that need through human-inspired learning.