NeoCognition $40M seed funding is putting fresh attention on a new wave of AI startups focused on making autonomous agents more reliable and useful in the real world.
The startup, founded by Ohio State professor Yu Su, has emerged from stealth with $40 million in seed funding. The round was co-led by Cambium Capital and Walden Catalyst Ventures, with participation from Vista Equity Partners and angel investors including Intel CEO Lip-Bu Tan and Databricks co-founder Ion Stoica.
Su described NeoCognition as a research lab building self-learning AI agents that can adapt and specialize the way humans do.
NeoCognition $40M Seed Funding Targets More Reliable AI Agents
According to Su, today’s AI agents still struggle with consistency.
He told TechCrunch that current tools, including systems from Anthropic’s Claude Code, OpenClaw and Perplexity computer tools, complete tasks as intended only about 50% of the time.
That level of reliability, in his view, means agents are not yet ready to operate as trusted independent workers.
Instead of building software that only handles one narrow task, NeoCognition wants to create agents that can learn on their own, improve over time and become experts in new environments.
Su said current agents are generalists, but human intelligence becomes powerful through specialization.
AI Agents Designed to Learn Like Humans
NeoCognition’s core idea is that AI systems should build internal models of specific environments, professions and workflows in the same way people learn new jobs or adapt to unfamiliar settings.
Su said humans quickly understand rules, relationships and consequences when entering a new field. He believes AI agents need that same ability to autonomously learn a “micro world” if they are going to become dependable experts.
That approach is meant to bridge the gap between broad-purpose assistants and highly capable domain specialists.
The company argues that while custom AI agents can already be trained for specific industries, those systems usually need heavy engineering for each use case.
NeoCognition says its goal is different: create generalist systems that can self-learn and specialize in any domain.
Enterprise Customers Are the Main Focus
NeoCognition plans to sell its agent systems primarily to enterprise customers, including established SaaS companies.
Those businesses could use the technology to build AI workers or upgrade existing products with more capable automation.
Su also highlighted the strategic value of Vista Equity Partners joining the round. As a major investor in software companies, Vista could help connect NeoCognition with businesses looking to modernize products using AI.
The company currently has around 15 employees, with most of the team holding PhDs.
As competition intensifies around autonomous software, NeoCognition $40M seed funding signals growing investor belief that the next major AI opportunity may come from agents that can truly learn and specialize.
Don’t miss out on our latest news—follow us for the latest AI news, breakthroughs