Bespoke Labs raises $40M to build training grounds for AI agents

Mountain View, California-based startup Bespoke Labs has raised $40M to build environments for training and testing AI agents, the company announced.

Why AI agents need training grounds

Today's AI agents can write code and answer questions, but according to Bespoke Labs, they still fall apart on long, messy tasks that require many sequential steps. The company builds environments that train and test agents on exactly these kinds of scenarios, rather than on one-off questions that current models already handle well.

The idea is that judging an agent by how it answers a single question isn't enough. A real-world task can stretch across dozens of steps, involve switching between tools, and require building up context along the way. That's where agents most often lose the thread and make mistakes.

• Bespoke Labs is based in Mountain View, California
• The company says it has raised $40M in total funding
• It was founded by Alex Dimakis and Mahesh Sathiamoorthy
• Its product: environments for training and testing AI agents on long, multi-step tasks

Who's behind Bespoke Labs

The company was founded by Alex Dimakis and Mahesh Sathiamoorthy. Bespoke Labs focuses on infrastructure for agent developers. Instead of every company building its own test scenarios to check agent reliability, Bespoke Labs offers ready-made environments that replicate the long, "messy" real-world tasks where today's agents most often stumble.

This approach saves time for teams deploying agents to production. Rather than building their own test scenarios from scratch and manually labeling where an agent succeeded and where it failed, they can rely on ready-made training and evaluation infrastructure.

What it means

The Bespoke Labs round is part of a broader trend. As AI agents move into real-world workflows, demand is shifting from the models themselves to the infrastructure for testing them: environments, benchmarks, and training grounds that show where an agent is genuinely reliable and where it breaks down over the long haul.

FAQ

How much did Bespoke Labs raise?

The startup raised $40M in total, the company announced.

Who founded Bespoke Labs?

The company was founded by Alex Dimakis and Mahesh Sathiamoorthy.

What does Bespoke Labs do?

The company builds environments for training and testing AI agents on long, multi-step tasks, where today's agents most often lose the thread and make mistakes.

Why are investors backing training environments rather than new models?

Because the models already handle one-off requests reasonably well. The main barrier to deploying agents in real work is reliability across long, multi-step scenarios, and that is exactly what these environments test.