The Token Economy: AI Infrastructure And The Future Of Compute

In this episode of FYI, Brett Winton and Charles Roberts sit down with Michael Stewart of M12, Microsoft’s venture fund, to examine how AI is reshaping software, infrastructure, and investment strategy. Michael explains how M12 identifies emerging signals across the AI ecosystem, from generative AI adoption to enterprise spending shifts. The conversation covers the evolution from SaaS (software as a service) to AI-driven models, pricing dynamics, and the growing importance of compute efficiency. They also discuss future opportunities in consumer AI, material science, and the need to rethink the data center as demand for AI accelerates.Key Points From This Episode: 00:00:00 Introduction to M12 and its role within Microsoft00:04:41 The current AI landscape and platform shift thesis00:09:47 The gap between technical capability and user adoption00:14:33 Compute constraints and the need for model efficiency00:20:55 The concept of “wholesale AI” and token-based enterprise models00:28:00 Uneven AI usage across users and organizations00:31:47 Capturing value: productivity gains vs. software pricing00:35:34 Investment themes: consumer AI and smart TV experiences00:38:58 AI in science and material discovery opportunities00:44:35 Introduction to the data center of the future00:51:35 Capital markets, AI demand, and infrastructure investment gaps00:55:53 Why current AI investment narratives may be backward-lookingEditing and post-production work for this episode was provided by The Podcast Consultant (https://thepodcastconsultant.com)