Cisco Bets on Artificial Intelligence to Overhaul Decades of Software
San Jose, Friday, 30 October 2026.
Cisco granted a 27-year-old engineer unlimited artificial intelligence access and a two-person team to rewrite 98% of its legacy software architecture within two years.
A High-Stakes Gamble on Gen Z Ingenuity
In a bold move that blurs the lines between corporate strategy and technological experimentation, Cisco Systems, Inc. (NASDAQ: CSCO) has tasked 27-year-old Padmanabhan Krishnamurthy—a former startup founder—with rewriting 98% of the company’s legacy codebase using unlimited AI tokens over the next 24 months [1][2][3][4]. The initiative, spearheaded by Cisco President and Chief Product Officer Jeetu Patel, represents a high-risk, high-reward bet on generative AI’s capacity to modernize decades of accumulated technical debt. Patel’s decision to entrust this monumental task to a young engineer reflects a broader corporate philosophy: challenging entrenched engineering practices by injecting fresh perspectives into the organization [2][3].
From Buried Talent to Corporate Linchpin
Krishnamurthy’s journey to this pivotal role began in 2023, when he co-founded EzDubs, a real-time translation startup that Cisco acquired in November 2025 [3][4]. Prior to his elevation, Krishnamurthy was reportedly ‘buried away in the organization doing something unstrategic,’ according to Patel [1]. However, his proposal to rewrite Cisco’s legacy code with a two-person team caught the executive’s attention, leading to Patel’s directive: ‘You can go, you can go nuts’—a phrase that underscored both the scope and the urgency of the task [1][2]. Patel’s approach is part of a deliberate strategy to foster ‘tension’ within teams, arguing that ‘fresh ideas get injected’ by challenging the status quo with younger, less conventional talent [2][3].
The Four-Week Proof and the Road Ahead
Krishnamurthy’s mandate includes a four-week demonstration period—beginning in late September 2026—to prove the feasibility of his approach before committing to the full 24-month overhaul [2][3]. This proof-of-concept phase is critical, as it will determine whether a two-person team armed with AI tools can generate the volume of code required to rewrite 98% of Cisco’s legacy architecture [4][alert! ‘The exact volume of legacy code to be rewritten is unspecified in available sources’]. Patel’s backing is rooted in the belief that AI-native engineers, unencumbered by traditional constraints, can accelerate modernization efforts far beyond what experienced teams might achieve [2][3].
Critiques and Cautions in the Age of AI-Driven Refactoring
Not all observers are convinced that AI regeneration alone constitutes true modernization. Critics argue that simply translating legacy code into newer languages—such as converting COBOL to Java—merely creates ‘JOBOL,’ where the syntax changes but the underlying architecture and data dependencies remain tangled [4]. True modernization, they contend, requires addressing six critical decisions: redefining system boundaries, migrating data structures, distinguishing business rules from bugs, eliminating obsolete code, planning traffic migration, and ensuring future maintainability [4]. Krishnamurthy’s plan assumes these challenges can be managed within the 24-month timeline, but the feasibility of such an ambitious undertaking without substantial architectural oversight remains a point of debate [alert! ‘No public statements from Cisco address these critiques directly’].
Broader Implications for Enterprise AI Adoption
Cisco’s experiment is being watched closely by enterprise leaders grappling with the same dilemma: how to leverage AI to reduce technical debt without sacrificing system reliability or scalability. The company’s team composition—roughly one-third experienced operators, one-third outside-in systems thinkers, and one-third AI-native founders and engineers—reflects a deliberate strategy to balance innovation with institutional knowledge [2][GPT]. Patel’s endorsement of generational friction as a catalyst for innovation aligns with trends highlighted by investors and executives, such as Thrive Capital’s Josh Kushner, who prefers hiring candidates with fewer than four years of experience, and Tastewise CEO Alon Chen, who hires entry-level employees to ‘avoid conventional thinking constraints’ [2][3]. For Cisco, the stakes are particularly high, as the success or failure of this initiative could redefine how Fortune 500 companies approach legacy system modernization in the AI era.