Cisco to rewrite 98% of its legacy code using AI

Cisco president and chief product officer Jeetu Patel has assigned 27-year-old Padmanabhan Krishnamurthy an ambitious project: refactor 98% of Cisco’s legacy code within 24 months, starting with a four-week proof of concept. Krishnamurthy joined Cisco when it acquired EzDubs, a translation startup he helped found, but Patel says he had been placed in an unstrategic role. After Krishnamurthy proposed that a two-person team could use AI to tackle the codebase, Patel gave him unlimited AI tokens and encouraged him to proceed. The plan reflects Patel’s belief that AI-native younger engineers and startup founders can question established practices and accelerate work that conventional teams might approach more cautiously.

Patel wants to bring more entrepreneurial early-career talent into Cisco, arguing that companies need fresh perspectives to challenge entrenched systems. He sees productive disagreement between younger hires, experienced employees, and seasoned operators as valuable, provided it focuses on ideas rather than personalities. The scale of the undertaking, however, raises questions about whether rewriting code is the hardest part: Cisco’s networking products require extensive testing, and regressions or defects could affect systems that underpin internet infrastructure. The project’s success will depend not just on AI’s ability to generate or refactor code, but also on validating changes and maintaining the reliability customers expect.