Threat actors using Frontier AI models are perhaps the most pivotal pacing cybersecurity threat of our generation. These threats recently prompted extraordinary governmental actions, including Presidential Executive Orders, export contol directives and public-private partnerships aimed at giving US government agencies, critical infrastructure owners, banks, insurance companies and others time to effectuate changes to their cybersecurity operations and lifecycle management practices before these models inevitably end up in nefarious hands. 

Broadcom was among the first software engineering organizations in the US permitted access to Frontier AI security models. What we learned testing those models against production code fundamentally changed how we think about cybersecurity, and we're not alone. This session will cover the Frontier AI threat landscape, how Broadcom is helping critical infrastructure organizations respond, and how the VMware by Broadcom capabilities the Dept of Air Force already owns can be optimized to enable 'speed as security' across three key attack vectors: virtual infrastructure, open-source containers, and application frameworks like Spring and Java.