John and Patrick Collison on Stripe's Growth, Agent Commerce, and the Future of Software
The a16z Show
Stripe’s data indicates a significant economic phase transition beginning in 2025, where new business cohorts are both larger and more productive, signaling that AI is finally delivering tangible value. This shift necessitates a fundamental redesign of financial infrastructure, as the rise of agentic commerce will require blockchains capable of processing billions of transactions per second—a capacity currently unavailable. Consequently, software development is moving away from traditional fixed-cost, mass-produced models toward bespoke, "freshly cooked" applications generated at the moment of use. Stripe founders John and Patrick Collison emphasize that these developments, including the incubation of high-throughput projects like Tempo, are driven by specific, real-world pain points rather than abstract market projections. This evolution reflects a broader trend where AI-driven productivity gains are becoming deeply embedded in the real economy, moving beyond initial hype to substantive operational change.
00:00Stripe Growth Trends and the Emergence of the AI Singularity
Stripe Growth Trends and the Emergence of the AI Singularity
Stripe experienced 34% growth last year, driven by a high-performing cohort of businesses that are increasingly adopting AI and stablecoin technologies. Data suggests that 2026 Q1 may mark the beginning of an AI-driven singularity, as the volume and performance of new businesses on the platform accelerate. Agentic commerce is identified as a major upcoming trend, necessitating the development of high-throughput blockchains capable of supporting billions of transactions per second to overcome current web and payment rail limitations.
09:28Infrastructure Development for Agentic Commerce and Bespoke Software Models
Infrastructure Development for Agentic Commerce and Bespoke Software Models
Executive skepticism regarding AI value is dismissed as inaccurate, as businesses are actively integrating AI to improve productivity and operational efficiency. Future software development is shifting from mass-produced, fixed-cost models to bespoke, "freshly cooked" applications created at the moment of use. This transition requires infrastructure-level work, such as building buyable product catalogs within AI apps and incubating projects like Tempo to solve blockchain congestion. Successful product innovation relies on solving specific, real-world founder pain points rather than relying on broad market sizing or traditional MBA-style business projections.
Sign in to continue reading, translating and more.
Open full episode in Podwise