Fleet

Reinforcement learning environments for frontier AI labs and enterprise agent workflows

AI 📍 New York, NY Founded 2024

Current Valuation

$750M

as of 2026-Q2

PRIVATE

Fleet builds high-fidelity reinforcement learning environments — practice arenas where AI agents repeatedly attempt multi-step enterprise workflows (claim processing, CRM updates, IDE tasks), fail safely, and learn before deployment. The company's Python SDK exposes versioned, deterministic environments programmatically; its Harbor evaluation framework runs thousands of parallel agent experiments. Annualized revenue jumped from roughly $1M at the end of 2025 to ~$60M by April 2026 on frontier-lab demand for post-training data, and Fleet is in talks to raise ~$50M at a ~$750M valuation led by Bain Capital Ventures with Sequoia, Menlo, and SV Angel participating.

Company Profile

CEO

Nicolai Ouporov

Founded

2024

New York, NY

Employees

~60

Total Raised

$15M

Est. Revenue

$60M (2026-Q1)

Growth

Annualized revenue scaled from ~$1M (2025) to ~$60M (April 2026) on frontier-lab demand for RL training environments

Last Round

Series A (in progress) — $50M (Apr 2026) · Lead: Bain Capital Ventures (in talks)

IPO Status

private

Founders & Key People

Nicolai Ouporov

Investors

Bain Capital Ventures · Sequoia Capital · Menlo Ventures · SV Angel

Products

Fleet Platform (versioned RL environments + SDK)
Harbor (agent evaluation framework)
Software-to-environment conversion frameworks (browsers, IDEs, EHRs, enterprise SaaS)

Competitors

Surge AI · Mercor · Turing · Scale AI · ServiceNow (WorkArena / BrowserGym)

AIRL EnvironmentsData LabelingFrontier Lab InfrastructureB2B