​Enterprise AI is no longer just a question of which model or vendor to choose. It's a question of control.

​Where does your data live? Who can inspect and reproduce system behavior? Can you change models without rebuilding the product? What happens when a vendor changes its pricing, policies, or performance—and what changes when the AI itself can adapt after deployment?

​Open models and increasingly capable agent systems give enterprises more leverage. They also move more responsibility in-house: security, reliability, evaluation, governance, and cost.

​The question is not whether to own everything. It is which control points are too important to rent.

​The conversation

• ​What it actually means to “own” your AI stack

• ​Which parts are worth bringing in-house—and which are not

• ​Where security, compliance, and data-residency requirements change the answer

• ​What tends to break when companies move beyond managed platforms

• ​How to evaluate the economics of building, buying, and renting

​We’ll begin with one candid panel, followed by dinner, drinks and continued discussion.

​Panelists

​Anjana Surin — Global Head, Visa Token Service Platform
Anjana leads the platform behind Visa Token Service, helping secure digital payments across wallets and e-commerce. She brings a global product perspective on identity, authentication, delegated authority, fraud, reliability, and introducing new technology into a trust network at scale.

David Hefter — AI Investment Strategist, BlackRock
David leads work across AI strategy, adoption, and governance at BlackRock, with a view into how a global institution evaluates platform dependence, operating risk, and the economics of enterprise AI.

Susanna Holt, PhD — CTO, OLI
A longtime enterprise technology leader and former Autodesk VP of Engineering, Susanna now leads technology at OLI, whose modeling and simulation systems support high-consequence decisions across chemicals, energy, mining, and industrial water. She brings an operator’s view of where probabilistic AI can—and cannot—be trusted.

Moderated by Kunal Bhatia, CEO and co-founder of Hexo Labs.

​The room

​Attendance is curated and capped. The room will include CIOs, CISOs, chief AI and risk officers, and the platform leaders directly responsible for enterprise AI decisions, along with a small number of founders and investors building the underlying stack.

​This is intended to be a working conversation among peers.

​Schedule

​5:30 PM — Arrival and drinks
6:30 PM — Panel discussion
7:15 PM — Dinner and continued conversation
8:30 PM — Close

​Hosted by

​Hexo Labs, the neolab behind SIA — an open-source framework for AI systems that improve through structured feedback. Hexo’s work raises one of the evening’s central questions: as AI systems become more adaptive, how do enterprises preserve auditability, rollback, and control?

​DevC is a founder-led venture fund and global collective of founders, operators, and early-stage GPs backing companies at the pre-seed and seed stages. DevC has invested in more than 150 companies over the past three years, including Hexo Labs.

​The Frontier Syndicate is a private venture network where frontier signal becomes conviction. Through high-trust rooms and conviction-led capital, it connects the researchers, builders, enterprise leaders, and investors shaping the next generation of technology.