Startup profile for Vals AI: what they do, why they are interesting, stage (series a), industries, and links to our funding articles. Part of the Venture Capital Tracker startup directory.

Startup profile · funding coverage

Vals AI startup funding profile

Independent evaluation platform and private benchmarks that measure whether AI models can do real professional work.

Stage

series a

Status

private

Coverage

full

Overview

Vals AI builds privately held test sets and evaluation infrastructure to score frontier models on lawyer, banker, engineer, and doctor work. On August 13, 2026 it announced a $40 million Series A at a $400 million valuation led by Andreessen Horowitz, with participation from existing investors 8VC, Pear VC, and BloombergBeta and new investors HRT Ventures and Next Ladder Ventures. The company said revenue grew 8x versus all of 2025, customers doubled, and headcount tripled in six months — treat those as company-reported. Alongside the round it launched Vals Smith (custom coding benchmarks from GitHub repos), frontier-risk benchmarks, and Vals Index 2.0.

Why Vals AI is interesting

August 13, 2026 Series A: $40M at $400M led by a16z, with 8VC returning. The bet is that labs and enterprises need a third-party scorekeeper as public benchmarks leak into training data — not another foundation model.

Product & use cases

Evaluation platform plus privately held professional-work benchmarks (finance, coding, legal) and Vals Smith for generating coding tests from GitHub repositories.

  • Labs citing independent scores in model cards
  • Enterprises choosing models for domain work
  • Policy and commerce teams measuring capability and cyber risk

Key facts

  • Series A (Aug 13, 2026): $40M at $400M led by a16z; 8VC, Pear VC, BloombergBeta returning (company blog)
  • Company-reported: 8x revenue vs 2025; 2x customers; 3x team in six months
  • Directory-linked: /fund/andreessen-horowitz, /fund/8vc

Fundraising history

Series A

2026-08 $40M Valuation: $400M

Source: https://www.vals.ai/blogs/series-a

Investors in our directory

Funds linked from Vals AI's profile — open a fund page for stage focus and related deal articles.

Competitive landscape

Edge: Private test sets and claimed model-card citations. Independence is the product; an a16z lead is a credibility tension to watch.

AI evaluation is splitting into public arenas, academic suites, and commercial independent testers. Vals is in the third bucket. Switching costs depend on whether private benches stay uncontaminated and whether buyers trust an a16z-backed vendor to score a16z-backed labs.

  • LMSYS / Chatbot Arena adjacent

    Public preference leaderboards, not private professional-work tests.

  • Stanford HELM adjacent

    Academic transparency evals; slower commercial packaging.

  • Lab-owned eval harnesses incumbent

    First-party scores used in marketing; conflicted by construction.

Notable stories

  • Company: results cited in model cards from OpenAI, Anthropic, Google, Meta, and xAI (Vals Series A post).
  • Thesis in the company’s words: trillions invested in generating intelligence, comparatively little in independent measurement.

Industries

AI & Machine Learning Developer Tools Enterprise SaaS

Related funding articles

Venture Capital Tracker pieces that cover Vals AI's financing or category context.

FAQs about Vals AI

Practical answers founders, operators, and investors typically search for.

Vals AI sells independent AI evaluation: private benchmarks and a platform to test whether models can do professional work in domains such as finance, coding, and law.
Series A (Aug 2026, $40M at $400M) led by Andreessen Horowitz (/fund/andreessen-horowitz). 8VC (/fund/8vc) returned from earlier rounds. Pear VC, BloombergBeta, HRT Ventures, and Next Ladder Ventures also participated (company).
Private Series A as of August 13, 2026.
No. It evaluates other companies’ models. It does not compete as a frontier-model trainer.
No. The 8x revenue claim is company-reported in the Series A post.
Public arenas measure chat preference. Vals says it uses private tests of professional work product — a different job for enterprise and government buyers.
San Francisco, California (public materials).
a16z-led round with 8VC on the cap table; independent eval is relevant to founders mapping AI infrastructure buyers.

Last updated: 2026-08-18