Halluminate Raised $30 Million for AI Sandboxes
The San Francisco startup provides managed testing environments and evaluation services for AI labs and browser agents.
Updated on Oct. 1, 2026 in Startups

Live Poll
Is now a good time to invest in AI infrastructure companies instead of model builders?
San Francisco-based startup Halluminate has secured $30 million in Series A funding to expand its managed sandbox environments for training AI agents. Founded in 2024, the company now maintains a total funding of $38.5 million.
Why it matters
As investment flows into infrastructure tools, Halluminate is positioning itself to support model builders with proprietary benchmarking and evaluation services. The company currently serves four US-based AI labs and two browser-based AI agent developers.
Halluminate has scaled its operations to nine employees while serving six active commercial clients. The company leverages managed sandbox environments to provide standardized evaluation metrics for AI model performance.
The players
Halluminate
A San Francisco-based AI infrastructure startup that builds managed sandbox environments and provides benchmarking services for model builders.
Oak HC/FT
A venture capital firm focused on the healthcare and fintech sectors that led the $30 million Series A funding round for Halluminate.
Y Combinator
A prominent startup accelerator that hosted Halluminate during its Summer 2025 cohort.
The details
Halluminate builds managed sandbox environments—isolated digital workspaces where software can be safely executed—to facilitate the training and testing of AI agents. The platform offers proprietary benchmarks and expert evaluation services to quantify model capabilities for labs and browser-based agent companies. These tools allow developers to isolate agent performance in controlled, replicable conditions.
Timeline
Halluminate was founded in 2024.
The startup participated in the Y Combinator Summer 2025 batch.
The $30 million Series A round closed on October 1, 2026.
The Tech Race
Halluminate's growth aligns with the broader push among Y Combinator-backed firms to standardize testing environments for large-scale model development. This infrastructure focus aims to solve evaluation bottlenecks that currently challenge US AI labs and agent-based software companies.
The startup's growth directly impacts developers at AI labs and browser-agent companies who need automated environments for model training. While currently focused on enterprise clients, these testing standards may eventually influence how developers build and deploy general-purpose AI agents.
The takeaway
Halluminate is scaling its infrastructure services to meet the demand for rigorous, controlled AI agent evaluation. Observers should track the expansion of the startup's customer base among major US labs to gauge the adoption of its proprietary benchmarking metrics.
Further reading
For more on the local venture landscape, see Startups.
Live Poll
Is now a good time to invest in AI infrastructure companies instead of model builders?







