Agentic AI Security: Broadcom's Infrastructure-Level Defense

Agentic AI is rapidly expanding the enterprise attack surface, shifting security deeper into the infrastructure layer. This means traditional perimeter defenses are no longer enough. Umesh Mahajan, from Broadcom, emphasizes that businesses must deploy lateral security now. He states that taking years to implement these changes is not an option. What's interesting is how Broadcom is building zero trust directly into cloud infrastructure. They offer an integrated software stack to close the security gaps created by piecemeal tools. The company is pushing security enforcement to the hypervisor level. This allows for high-volume inspection, like 75 terabits per vCenter cluster for firewalling, without adding significant delays. The bottom line is that with AI workloads, low latency is critical, making integrated, infrastructure-level security essential for protecting against new threats.