Noeri Raises $110M to Build Self-Improving Enterprise AI
Noeri launched with $110M led by Menlo Ventures to build self-improving enterprise AI. Stanford professors Yejin Choi and Carlos Guestrin lead the Palo Alto and Zurich lab.
By Daniel Okafor
3 min read

What's News
- Noeri raised $110 million announced November 16, 2026, with Menlo Ventures leading the round.
- Stanford professors Yejin Choi and Carlos Guestrin founded Noeri in 2026 and serve as co-CEOs.
- Cisco Investments, NVIDIA and Snowflake joined as strategic investors.
- Every proposed system change reaches a human for approval through a process similar to code review.
- Guestrin previously co-founded Turi, which Apple acquired in 2016.
By Daniel Okafor
Noeri launched on November 16, 2026 with $110 million in funding, led by Menlo Ventures, to build self-improving AI for enterprises and scientific discovery. Madrona Venture Group, Kindred Ventures, Wing Venture Capital and NEA joined the round. Cisco Investments, NVIDIA and Snowflake signed on as strategic investors.
The founders are Stanford computer science professors Yejin Choi and Carlos Guestrin, who serve as co-CEOs. The company is based in Palo Alto and runs a second office in Zurich. It is hiring researchers and engineers in both locations.
What problem does Noeri attack?
Enterprises face a familiar trade-off. They can pay more for the most capable general-purpose model, or accept weaker results from a cheaper one. Noeri says its system breaks that choice apart. It studies an organization's data and workflows, then improves the surrounding software, skills, instructions, tools and the choice of model for each step. It will tune a model's underlying parameters when that pays off.
The product overview describes agents that learn from tasks, failures, corrections and real production traffic. No large labeled dataset is required. The agents refine their own skills, tools and orchestration. They can build new tools when existing ones fall short, and they train models on the customer's actual work.
The measurement standard differs from the industry norm. Noeri tracks quality and cost on the customer's own traffic and metrics, not public benchmarks. That makes every claimed gain specific to one deployment rather than to a leaderboard.
How does human control work?
Every proposed change goes to a person for approval. The review process resembles code review, and the system explains each change before it lands. The system runs on both closed and open-weight models and adapts as new models arrive.
Ownership stays with the customer. The customer keeps its data and owns the workflows and models built through the system, though applicable third-party model licenses still apply.
Why smaller models for science?
Noeri plans to point the same machinery at scientific discovery. It wants systems that form hypotheses, run experiments and learn from results in work on molecules, proteins and complex technical systems.
The company's argument is economic. Open-ended discovery strains the economics of large models because sustained search consumes far more compute than a single answer. Smaller specialized models are Noeri's proposed way to make that search affordable.
Its research agenda covers four areas:
- automating AI development;
- specializing efficiently from limited data;
- continual learning without forgetting;
- discovery through search, reasoning and teams of models.
Those priorities build on the founders' prior work in test-time training, self-distillation and agentic orchestration, where models learn after deployment, learn from their own mistakes, or coordinate other models and tools.
What do the founders believe?
In a vision statement, Choi and Guestrin argue that intelligence should be distributed across many specialized, customer-owned systems instead of concentrated in one centralized superintelligence. They also make a cost case: models sized to a task are cheaper and less energy-intensive, and purpose-built systems are easier to specify, test and supervise.
Who is behind the company?
Guestrin co-founded Turi, the machine-learning platform Apple acquired in 2016. He later built and led Apple's central machine-learning and AI organization. Choi is a MacArthur Fellow whose research focuses on making smaller AI models more capable. She was previously a professor at the University of Washington and a senior director at the Allen Institute for AI.
Noeri says its first proof-of-concept partnerships span enterprise and scientific domains. It is also working with strategic partners across industries, including some of its investors. The partner group is deliberately varied, so the technology learns to generalize before specializing for individual customers.
The immediate test is concrete: whether those partnerships turn Noeri's research into repeatable improvements on customers' own measures of quality and cost, while preserving the human approval and ownership controls at the center of the product.
Source: unite.ai
More from Daniel Okafor
Show full bio
Correspondent covering money, funding and market shifts around entrepreneurship, startups and venture funding at Business Bearings.
2,189 articles