A new startup called Sycamore just raised $65 million to solve the problem that every business running AI agents eventually hits: how do you govern dozens of autonomous agents without losing control of your operations?

The round was announced March 30, 2026, and the investor list reads like a greatest-hits of enterprise tech. Coatue and Lightspeed Venture Partners led the round. Angels include Databricks CEO Ali Ghodsi, former OpenAI Chief Scientist Bob McGrew, Intel CEO Lip Bu-Tan, Palo Alto Networks President BJ Jenkins, and AI researcher François Chollet. This is not a speculative bet from generalist VCs. These are operators and builders who understand exactly what enterprises are running into right now.

What Sycamore is building

Sycamore’s founder is Sri Viswanath, former Chief Technology Officer at Atlassian. He knows enterprise software at scale, and the problem he is tackling is one the industry has quietly been circling for months.

The company describes itself as building a “trusted agent operating system” for enterprises. Practically, that means a centralized layer where businesses can discover, deploy, monitor, and control fleets of AI agents running across their organization.

The specific problem Sycamore calls “operational gravity.” Here is what that looks like in practice: a business deploys an AI agent for customer follow-up, another for document processing, another for scheduling. Each one works in isolation. There is no central view of what they are doing, no way to enforce company policies across them, and no way to ensure they are staying within security boundaries.

As the number of agents grows, so does the risk of conflicts, errors, and compliance gaps. Sycamore’s platform is designed to be the control layer that sits above all of it.

The system offers tiered trust controls, meaning different agents can be granted different levels of autonomy based on the stakes involved. An agent handling internal scheduling can operate with minimal oversight. An agent making procurement decisions needs tighter controls. Sycamore lets you define those levels and enforce them systematically.

Why this funding round matters

The caliber and composition of this investor group tells you something important: the enterprise AI infrastructure layer is now considered a critical investment category, not a speculative one.

Every major AI lab founder on the cap table has watched enterprises struggle with the same problem. You can deploy a single agent. You can prove it works. But scaling from one agent to fifty agents across a complex organization requires infrastructure that does not yet exist at most businesses.

The fact that former OpenAI Chief Scientist Bob McGrew is personally writing a check into this tells you the AI labs themselves understand this is the missing layer.

Sycamore was founded at the end of 2025 and is raising $65 million before it has shipped a product publicly. That is a very strong signal of where enterprise AI infrastructure investment is going in 2026.

What this means for business

If you are in the early stages of deploying AI agents in your business, this story is directly relevant, even if you have never heard of Sycamore.

The governance challenge is real and it is coming for every business that takes AI agents seriously. Most businesses start with one agent, maybe two. The ROI is clear, so they add more. At some point, the question stops being “does this agent work?” and starts being “how do I know all of my agents are behaving correctly, not contradicting each other, and staying within the boundaries I have set?”

That is not a technology question. That is an operations question. And most businesses are not ready for it.

A few things to consider as you plan your own AI agent deployment:

Start with clear scope definitions. Every agent you deploy should have a defined scope of authority. What can it do without human approval? What requires sign-off? Documenting this now makes governance easier as you scale.

Audit trails matter from day one. You want to know what every agent did, when, and why. Building this habit early, even with simple logging, means you have a foundation to build better governance on as your agent fleet grows.

Think in systems, not individual tools. The businesses that deploy AI agents successfully do not think about them as individual software tools. They think about them as an operational layer that needs to be managed, optimized, and governed just like any other part of their operations.

The size of this funding round reflects how many enterprises are now reaching the point where governance is the bottleneck. The pilot phase is over. Production deployment is here. And the infrastructure to support it at scale is what the investment community is now building.

Sycamore’s launch is a signal, not just a funding announcement. It tells you where the industry is right now: past the “can AI agents work?” question and squarely into “how do we run them at enterprise scale?”

For businesses that are serious about AI agents, the answer is not to wait for platforms like Sycamore to mature before you start. It is to build good habits now, deploy thoughtfully, and position yourself to take advantage of this infrastructure as it arrives.

The window to get ahead on operational AI is still open. But the investment activity in enterprise AI infrastructure suggests it is narrowing.

If this is the kind of problem agents can help with, the free Working With Claude field guide is the practical next step. Thirty-two pages, no fluff. Get the free guide.