How Jeetu Patel Runs Cisco Like the World’s Largest Startup

Patel describes Cisco as a picks-and-shovels provider for the AI era, building infrastructure that helps AI run securely. He says his role is to set direction, hire well, and choose markets, while giving teams credit for the product work. His leadership team follows a rule of thirds: Cisco operators, outside leaders, and founders or CEOs from acquired companies.

Patel says Cisco is in a high-stakes period and needs to combine startup speed with the reach of a large company. He sees two major needs: more networking capacity as agent use grows, and security tools to monitor agents and prevent harmful actions. Cisco’s agentic security platform includes visibility into models and agents, non-human identity, red teaming, and runtime guardrails.

Patel says Cisco introduced AI to its 32,000 engineers and the wider company by giving employees tools and unlimited tokens, while making clear that they were expected to use them. He argues AI should help teams pursue work they could not previously tackle, rather than simply reduce headcount. A skeptical engineering leader changed his view after a team refactored roughly half a million lines of code in two weeks.

Patel describes different planning cycles across Cisco’s stack: about five years for silicon, 18 to 24 months for hardware, and as little as weeks for models and apps. He says Cisco’s networking connects GPUs within racks, across data center rows, and between sites separated by hundreds of kilometers. AI networking orders grew from $2.3 billion in the first year to $9.3 billion two years later.

Patel recounts leaving India at 19, taking over a company as an intern, and running it for 17 years before moving to EMC and then Box. He says he stayed too long in a business that did not match his ambition, partly because of ego, and learned a great deal from mentor Rick at EMC. His advice to founders is to think boldly, build durable relationships, improve steadily, and spend time on work they will value later.