Oso and Monad Convene Security Leaders at AI Week New York to Define the Next Layer of AI Infrastructure

Autonomous systems are moving faster than their control layers can keep up. Not in benchmark wars or model releases, but in permission systems, infrastructure decisions, and the uncomfortable silence that follows a simple question: what exactly are these agents allowed to touch? Teams are deploying autonomous systems deeper into production while governance models are still catching up. The gap between capability and control is widening, and it’s starting to define the next phase of the startup ecosystem.

That gap is exactly where the Security Leaders Breakfast at #AIWeekNY plants its flag on May 13 in New York City during AI Week New York 2026. Oso and Monad are not building another crowded panel stuffed with recycled talking points and safe applause lines. They are building a closed-door breakfast for security leaders dealing with the operational reality of AI agent adoption. No presentations. No pitches. Just practitioners comparing notes before the rest of the market catches up to the problem.

The room is being shaped by Meghan Gill, Valerie Zargarpur, Darwin Salazar, and Pulse NYC, with Christian Almenar also publicly tied to the event alongside other confirmed attendees. Meghan Gill brings the experience of helping scale MongoDB from its earliest chapters into enterprise infrastructure gravity, now focused on go-to-market and community efforts at Oso. Graham Neray, CEO of Oso, has positioned the company around authorization systems that developers can actually maintain as complexity compounds. Christian Almenar, CEO of Monad, operates from the visibility side of the equation, helping security teams normalize and operationalize security data across fragmented systems. Together, Oso and Monad represent two converging realities inside the startup ecosystem: control and observability now move together or not at all.

The conversations inside this breakfast are pointed directly at the friction operators are feeling right now. Designing least-privilege access for autonomous agents. Keeping authorization logic maintainable as infrastructure sprawls. Securing systems without slowing product velocity into bureaucracy. Every company pushing AI deeper into production eventually runs into the same wall: scale exposes weak permissions faster than any penetration test ever could.

Pulse NYC gives the gathering additional gravity. AI Week New York has become one of the city’s strongest signals that the center of AI conversation is shifting from spectacle toward implementation. Founders, engineers, executives, investors, and policy voices move through dozens of events during the week, but smaller curated rooms like this are where the most valuable information usually changes hands. Not because someone is selling certainty, but because experienced operators finally admit what is breaking.

That is what makes this breakfast matter. The market spent the last cycle obsessed with what AI systems could generate. This cycle is becoming about what they can access, what they can execute, and who carries accountability when autonomous systems move faster than governance structures. Inside the startup ecosystem, the companies that solve that tension cleanly will not just move faster. They will be trusted faster.