Startup profile for Noeri: latest funding, latest known valuation, total disclosed funding, investors, status, sources, and why the company is interesting. Part of the Venture Capital Tracker startup directory.

Startup profile · funding coverage

Noeri funding, valuation and investors

Enterprise AI company building continuously improving domain-specific agents, model routing and customer-owned AI workflows.

Verified company research · funding, investors and latest developments

Keep track of Noeri

Save this profile to your VCT watchlist for a quick return.

View watchlist

Funding, valuation & investors

Answer-first snapshot

Latest funding

Funding (round stage undisclosed)

$110M · November 16, 2026

Latest known valuation

Not publicly disclosed

Total disclosed equity funding

$110M

Excludes debt, grants, acquisitions, secondaries, and IPO proceeds.

Current status

private

Stage not recorded · Palo Alto, California, United States

Investors in latest funding

Lead: Menlo Ventures

Other: Madrona Venture Group , Kindred Ventures , Wing Venture Capital , NEA , Cisco Investments , NVIDIA , Snowflake

Sources: latest funding · madrona.com Last verified: 2026-11-17

Overview

Co-founded by Stanford researchers Carlos Guestrin and Yejin Choi, Noeri proposes a continuously improving enterprise-AI system that learns from task outcomes, runs controlled experiments on tool use and agent workflows, and proposes changes for human approval. It aims to optimize quality and compute costs on a company's own data; the vendor says enterprises retain their resulting AI assets, subject to underlying model licensing.

The publicly disclosed $110M financing was announced 16 November 2026; its exact round label, valuation, ARR, deployment count, public pricing and audited improvement rates are not disclosed. Enterprise buyers should validate data-use terms, experiment controls, regressions, rollback, ownership and integration scope.

Why Noeri is interesting

The $110M November 16 financing led by Menlo Ventures is a large bet on improving deployed enterprise AI rather than just training a new frontier general-purpose model. The investor and product sources describe a human-approved system for modifying agent tools, instructions and sometimes models.

Product & use cases

Illustrative enterprise buyer journey: (1) select a bounded existing agent workflow and performance baseline; (2) connect task logs, outcomes and evaluation signals; (3) Noeri identifies prompt/tool/model-routing and potential model-weight improvements; (4) reviewers inspect suggested changes and authorize deployment; (5) run experiments on representative traffic, monitor errors and cost; (6) maintain versioned approvals, rollback and ownership checks. Vendor promises ongoing improvement; commercial benchmark evidence is not yet independently established.

  • Reduce quality failures in existing enterprise AI agents using real execution data and human approval.
  • Customize domain-specific models where generic frontier models are costly or inaccurate.
  • Build repeatable evaluation, routing and tool-evolution loops for internal AI systems.

Key facts

  • $110M announced 16 November 2026, led by Menlo Ventures.
  • Participating investors include Madrona, Kindred, Wing, NEA, Cisco Investments, NVIDIA and Snowflake.
  • Co-founders and co-CEOs Carlos Guestrin and Yejin Choi are Stanford AI researchers.
  • Round stage and valuation not identified in issuer-issued announcement; do not label the entire amount a Series A or Seed without original confirmation.

Funding history (newest first)

Funding (round stage undisclosed)

2026-11-16 $110M
  • Menlo Ventures (lead)
  • Madrona Venture Group (participant)
  • Kindred Ventures (participant)
  • Wing Venture Capital (participant)
  • NEA (participant)
  • Cisco Investments (participant)
  • NVIDIA (participant)
  • Snowflake (participant)

Source: Read source · madrona.com

Competitive landscape

An improvement learned from one customer workflow is not general intelligence: eval leakage, nonstationarity, costly online experiments, confidential-data transfer and rollback governance must be validated. Public model and product claims come from issuer/investors, not independent production testing. Price and actual cost-per-successful-task improvement are undisclosed.

  • LangSmith adjacent

    Agent traces and evaluation tooling; automation of model-level self-improvement is a separate capability to validate.

  • Weights & Biases Weave adjacent

    Tracing and evaluation for AI application iteration.

  • Braintrust adjacent

    Evals and experiment management; review differences in automatic optimization scope.

  • Predibase alternative

    Specialized fine-tuning and serving rather than entire agent workflow self-improvement.

  • In-house ML platform incumbent

    Retains control but requires staff for experiment design, approvals, deployment and monitoring.

Market / IPO context

Editorial / static context — not a live quote.

By Venture Capital Tracker

Last updated:

Editorial note: AI tools assisted with research, structure, or drafting. Venture Capital Tracker retains human editorial responsibility for factual accuracy, relevance, and source quality before publication.