Broadcom: Agentic AI expands attack surface, pushing security into hypervisor

Mahajan spoke with theCUBE's John Furrier during an exclusive broadcast at the event. He said the pace of change means enterprises can no longer postpone security decisions. "This is the time where you can't put off security any longer," Mahajan said. "Oh, I got a perimeter firewall. I'm good.' No, no, no – not good. It can be bypassed. Now the security gurus or experts are saying, 'No, you can't take two years, three years. You have to deploy lateral security.'"

Broadcom's response, Mahajan explained, is an integrated software stack delivered through its vDefend and Avi Load Balancer product lines. He noted that customers have bought multiple security products over the years but cannot stitch them together. "It's like buying Swiss cheese. Yeah, you have pieces of security, but you have plenty of holes which people can drive through."

Scale is another constraint. AI workloads generate heavy east-west traffic and punish any inspection step that adds delay, so Broadcom has pushed enforcement into the hypervisor rather than a separate appliance tier. The company says it is handling firewalling at 75 terabits per vCenter cluster and intrusion detection and prevention at 17 terabits per vCenter cluster, while keeping latency low for AI workloads.

Mahajan said visibility is also essential because agents and Model Context Protocol services are transient. Administrators need a real-time picture of what is authorized and what is shadow IT before they can quarantine anything. He said this visibility work now lands alongside private cloud modernization programs, and bolting on protection later, once production AI traffic is already running, is the failure mode executives are trying to avoid.

"It has to be at the infrastructure level; it has to be at scale," Mahajan said. "Otherwise, when are you going to do it? Two years from now, by that time you'll be compromised."