As AI moves into everyday work, two security gaps have emerged. Organizations want to use powerful cloud-based models without exposing sensitive data and intellectual property to third-party providers. And autonomous AI agents now run on employee endpoints, calling tools and accessing critical systems largely outside the view of traditional security controls. These challenges are driving a new layer of AI security: verifiable, hardware-backed privacy for AI workloads, and runtime visibility and response for agents in action. In this webcast, venture investors and CISOs will discuss how these approaches fit together and what security leaders should prioritize next.

Key discussion points:

• Why hardware-backed, verifiable privacy is replacing contractual promises as the standard for sensitive AI workloads.
• How agent sprawl and over-privileged tool access are expanding the attack surface, and what effective runtime oversight looks like.
• Which AI security layers investors expect to become core enterprise infrastructure, and what separates the startups likely to lead.