Skip to content

Noeri lands $110M to build "living" AI systems that self-improve—without relying on a sing

·StartupReader editorial desk· 3 min readReviewed by our editors
FUNDING. $110M raised by Noeri, self-improving AI systems. led by Menlo Ventures, Madrona, NEA, Snowflake, Cisco Ventures.

Noeri, a startup founded by two Stanford AI researchers, has raised $110 million to help enterprises build AI systems that learn and adapt over time—without locking them into a single proprietary model. The round, led by Menlo Ventures with participation from Madrona, NEA, Snowflake, and Cisco Ventures, reflects a broader industry push for cost-effective, self-optimizing AI that doesn’t depend on a handful of dominant providers.

The company’s core product is a "living" AI system that studies a business’s data, workflows, and processes, then suggests incremental improvements—like adjusting tools, refining instructions, or swapping out models—while keeping humans in the loop. It’s designed to address a common dilemma: companies often face a choice between expensive, high-performance models and cheaper, less capable alternatives. Noeri aims to bridge that gap by letting businesses fine-tune their own workflows over time.

The approach stems from research by co-founders Yejin Choi and Carlos Guestrin, both former Stanford professors with decades of experience in AI. Choi’s work on "agentic orchestration" demonstrates that smaller, specialized AI agents can sometimes outperform larger models at a fraction of the cost. Guestrin, meanwhile, developed a "self-distillation" technique where models improve by learning from their own mistakes—a process that forms the foundation of Noeri’s systems.

The startup’s name—*Noeri*, meaning "intelligence in the plural"—reflects its broader vision. Choi and Guestrin have long argued that the future of AI shouldn’t be controlled by a few centralized providers. Their critique mirrors concerns about over-reliance on a single vendor, where businesses risk dependency on terms that can change unpredictably or models that can’t be fully integrated into their operations.

This perspective aligns with a growing trend in AI startups. Last year, Thinking Machines Labs launched *Tinker*, a service for fine-tuning models, while Applied Compute helps clients optimize AI systems through embedded engineers. Noeri’s innovation lies in automating that optimization, allowing businesses to refine their workflows without deep custom engineering.

The funding round highlights Noeri’s immediate momentum. Menlo Ventures’ Venky Ganesan noted the rarity of combining cutting-edge research with practical deployment—a nod to Guestrin’s background. He co-founded GraphLab (later acquired by Apple) and created XGBoost, a widely used machine learning tool. Choi, meanwhile, led Nvidia’s common-sense reasoning team before returning to Stanford, where she helped launch Noeri earlier this year.

Noeri isn’t alone in challenging the dominance of monolithic AI models. When we covered *Dulo* in late October, Sebastian Thrun’s startup took a different approach by building domain-specific AI models for engineering tasks. But where Dulo focuses on niche applications, Noeri aims to be a general-purpose "operating system" for enterprise AI, adaptable across industries.

The involvement of investors like Snowflake and Cisco suggests Noeri is targeting sectors where AI integration is critical but proprietary models are impractical. Choi told Bloomberg the company is working with partners across multiple fields, ensuring its systems are tested on diverse workflows. While Noeri hasn’t disclosed customer names, some of its backers may be early adopters.

What’s next will depend on whether Noeri can prove its self-improving systems deliver real efficiency gains—and whether enterprises will embrace decentralized AI over the convenience of off-the-shelf solutions. The bet here isn’t just on AI, but on a fundamental shift in how companies build and control their own intelligence.

Sources: siliconangle.com

“This funding signals a shift toward decentralized AI workflows, where companies won’t need to bet everything on a single vendor’s model—and investors are backing the bet.”
— StartupReader
ShareLinkedInXWhatsApp

One of the funding rounds reported in the week of 16–22 November 2026. See every round that week.

Read the original reporting

The outlets below did the original reporting.

Related briefs

This brief was drafted automatically from the sources above and published under our editorial policy. Spotted an error? Tell us.