Stripe’s John Collison: AI Agents Will Rewire the Internet

Strategic Shifts and Market Performance

• Stripe processed $1.9 trillion in payment volume in 2025, representing a 34% year-over-year increase.
• The company views 2026 (specifically January 4) as the start of a new "Singularity epoch," marking the beginning of observable AI recursive self-improvement and rapid capability gains.
• New firm creation rates on the Stripe platform have reached the highest levels in recent memory, driven by the reduced cost and increased accessibility of AI-powered tools for founding.
• Stripe predicts a significant increase in cybersecurity breaches over the next five years, as AI-empowered offensive capabilities currently outpace defensive adoption rates across the industry.
• Cybersecurity risk profiles are bifurcating, with startups in better positions to adopt modern AI-driven security due to homogenous, agile tech stacks compared to legacy enterprises reliant on unsupported infrastructure.

Agentic Commerce and Market Dynamics

• The fundamental value proposition of commerce is shifting from human-filled web forms to stateful AI agents capable of autonomous execution and computer manipulation.
• "Computer use" is positioned as the ultimate backward compatibility layer, allowing AI agents to interact with existing digital infrastructure (e.g., government portals like the DMV) without requiring native API integrations or MCP server adoption by legacy institutions.
• AI-driven product discovery is acting as a new point of aggregation, potentially displacing traditional aggregators and favoring the "long tail" of niche businesses that rely on AI to surface high-quality, well-reviewed products.
• The commerce landscape is undergoing a "second-order" transformation similar to the rise of Facebook/Instagram ad targeting, where native adoption of AI acquisition tools allows new generations of companies to displace incumbents less adaptable to algorithmic distribution.
• Existing top-tier businesses face pressure to adapt their merchant discovery models, as AI agents may bypass traditional hierarchies to favor products with superior real-world performance data.

Organizational Transformation and Acquisitions

• Stripe's acquisition strategy has pivoted to cover three core AI-era needs: crypto-native culture (Bridge, Privy), usage-based billing mechanics necessitated by AI inference costs (Metronome), and multi-model routing infrastructure (OpenRouter).
• OpenRouter was acquired based on the thesis that enterprise AI workflows will require simultaneous usage of multiple models, with dynamic routing decisions based on user lifetime value and specific workload requirements.
• Stripe is rapidly adapting its internal developer experience (DX) from human-centric APIs to "AI ergonomics," recognizing that coding agents require different integration patterns and error-handling capabilities than human developers.
• Internal tooling at Stripe has evolved to address the "jagged frontier" of AI performance by creating custom applications that ingest internal data with strict privacy guardrails, enabling sales and legal teams to build personalized AI agents for their specific workloads.
• The "build versus buy" decision framework at Stripe now heavily weighs market velocity; in rapidly evolving sectors like token routing, acquiring established players is preferred over internal development to avoid the risk of the solution becoming obsolete before launch.

Future Outlook and Operational Philosophy

• The primary emerging technical challenge is the development of new "bot detection" equivalents capable of distinguishing between legitimate AI agents and malicious actors, as traditional bot prevention methods are becoming ineffective against AI-driven traffic.
• John Collison emphasizes a "long-term game" operating philosophy, citing Stripe's 17-year trajectory of consistent, non-exponential growth as evidence that "stick-with-it-ness" is a critical differentiator against hype cycles.
• Stripe's first operating principle remains "users first," requiring product development to start with customer utility rather than internal revenue targets or speculative market trends.
• To mitigate "AI psychosis" and maintain focus, executives prioritize analog interventions, including spending time in nature and disconnecting from digital inputs, rather than consuming academic papers or industry news.
• The company anticipates that AI will become the primary differentiator in business success, with the line between "AI companies" and "traditional companies" blurring as legacy firms integrate AI into core operational workflows.