20VC: Nikesh Arora on AI Token Costs, Memory Moats & Why Enterprise AI Lags

AI’s next battleground may be less about model size and more about memory, pricing, and trust. This minutes-long summary condenses Harry Stebbings’ full conversation with Nikesh Arora, CEO of Palo Alto Networks, into the key ideas on enterprise AI, token economics, cybersecurity, and leadership. Arora shares how he went from arriving in the U.S. with two suitcases and $200 to running one of tech’s biggest security companies, and explains why his mindset is focused on constant improvement rather than blame. He breaks down why consumer AI can tolerate mistakes while enterprise AI cannot, why long-term memory could become the real moat, and how token costs are likely to fall over time. The episode also explores AI’s impact on jobs, the future of cybersecurity, and why businesses should integrate AI carefully rather than “AI wash” their products. Listen now to get the key ideas in minutes.