Ed Sim, boldstart ventures | The AI Security Summit

In this AI Security Summit interview from San Francisco, Boldstart Ventures founder & general partner Ed Sim joins theCUBE’s Jackie McGuire to break down why securing AI must start with the code and extend across models, agents and infrastructure. Sim revisits Boldstart’s early bet on Protect AI – funded pre-ChatGPT and later acquired by Palo Alto Networks for $700M+ – as an early proof point that AI security is real and enterprise-scale. He details the evolution from model registries and model scanning (e.g., checking downloaded models for vulnerabilities) to capabilities such as automated red-teaming and agent security to meet “one-vendor” buyer demands. The conversation spotlights Snyk’s newly announced agent product, Eva, and why scanning source code first is foundational: secrets in code, MCP servers and unknown endpoints (often numbered in the hundreds) are where risk surfaces before models ever run. The discussion dives into the “AI security chasm” between rapid AI adoption and security readiness: AI-assisted coding has gone from experiments to wall-to-wall across the Fortune 500, driving far more (and larger) commits with fewer pull requests – amplifying maintenance and governance challenges. Sim advocates guardrails like spec-driven development to balance determinism and AI-assisted generation. He outlines why enterprises are leaning into hybrid computing – cost, data sovereignty and accuracy – using open-source models (e.g., Qwen) on-prem and orchestrating with Kubernetes. As agentic systems scale, identity must be rethought: agents are ephemeral, workflow-spanning and should have limited, revocable access, with their own audit logs – an area teams like Keycard are tackling. Looking ahead, Sim flags a new hardware wave (beyond GPU single-threading) and shares how startups such as Grepr.ai are cutting observability storage costs by ~90%, while a stealth effort targets the memory–GPU bottleneck with PCIe memory expansion and optical interconnects – aiming at 5–10× performance gains. His advice to founders: this is the biggest platform shift of our careers; start AI-native, use agents, stay lean and build security into the DNA from day one. Find more SiliconANGLE news and analysis https://siliconangle.com/ Follow theCUBE's wall-to-wall event coverage https://siliconangle.com/events/ Learn about the latest theCUBE events https://www.thecube.net/ 00:00 - Intro 00:06 - Foundations and Challenges of AI Security 02:41 - Securing Modern AI: Navigating Code and Complexity 05:19 - Integrating AI: From Specifications to Agency 07:51 - Evolving Dynamics in Cloud and Hybrid Computing 10:20 - Empowering AI: Navigating Data Ownership, Personalization, and Security 13:46 - The Next Wave of Hardware Innovation 16:15 - Building Foundations: Infrastructure, Data, and the Future of AI Entrepreneurship #theCUBE #SnykAISecurity #theCUBEresearch #BoldstartVentures #Snyk #ProtectAI #PaloAltoNetworks