Halluminate Raises $30M Series A Led by Oak HC/FT for AI Agent Training Environments

Halluminate Closes $30M Series A

Halluminate, a San Francisco-based startup building realistic sandbox environments and proprietary datasets for training and evaluating computer-use and browser-use agents, has raised a $30 million Series A round. The funding was led by Oak HC/FT.

Additional participation came from Orange Collective, Heavybit, Y Combinator, and FT Partners.

What Halluminate Does

Halluminate operates at the intersection of AI agent development and data infrastructure. The company builds realistic sandbox environments alongside proprietary datasets that allow AI labs and enterprises to train and evaluate computer-use and browser-use agents — systems designed to interact with software interfaces the way a human would.

These environments serve a dual purpose: they provide the training ground where agents learn to navigate real software workflows, and they offer the evaluation infrastructure needed to measure whether those agents perform reliably before deployment.

A Sharpened Focus on Financial Services

Since its founding, Halluminate has narrowed its commercial focus toward reinforcement learning (RL) environments targeting economically valuable financial-services workflows. Specifically, the company has oriented its product around investment banking, private equity, and consulting — domains where computer-use agents could take on structured, high-value tasks that currently consume significant human hours.

That positioning places Halluminate in a category of infrastructure providers selling picks and shovels to both frontier AI labs and enterprises looking to deploy agents in production settings.

Company Background

Halluminate was founded in 2024 by Cornell computer science alumni Jerry Wu and Wyatt Marshall. The company is headquartered in San Francisco and participated in Y Combinator's Summer 2025 batch.

Y Combinator's continued participation in this Series A follows its earlier backing through the accelerator program.

The Investor Group

The round brings together a mix of institutional and specialist investors:

• Oak HC/FT — lead investor
• Orange Collective
• Heavybit
• Y Combinator
• FT Partners

The presence of FT Partners, a firm known for financial-technology and financial-services advisory and investment work, aligns with Halluminate's stated focus on banking, private equity, and consulting workflows.

Why This Round Matters

The market for computer-use and browser-use agents has become one of the more competitive fronts in applied AI. Labs and enterprises alike need environments where agents can be tested against realistic software conditions — not toy benchmarks — before those agents are trusted with consequential tasks.

Halluminate's bet is that proprietary datasets and sandbox environments tailored to specific, economically valuable verticals will be more defensible than generic tooling. By concentrating on financial-services workflows, the company targets buyers with clear ROI calculations and willingness to pay for reliability.

What to Watch

With $30 million in new capital, Halluminate is positioned to expand its environment library and dataset coverage. Key questions for the coming period include how quickly the company can deepen its financial-services verticals, whether it extends into adjacent enterprise domains, and how it differentiates against other providers racing to supply agent-training infrastructure.

For AI labs and enterprises evaluating computer-use agents, Halluminate's trajectory is one to track — particularly as RL environments become a bottleneck in agent development pipelines.

Summary

• Company: Halluminate
• Round: Series A
• Amount: $30,000,000 USD
• Lead Investor: Oak HC/FT
• Other Investors: Orange Collective, Heavybit, Y Combinator, FT Partners
• Founded: 2024
• Founders: Jerry Wu, Wyatt Marshall
• Headquarters: San Francisco
• Accelerator: Y Combinator (Summer 2025)
• Focus: Sandbox environments and proprietary datasets for computer-use and browser-use agents; RL environments for financial-services workflows