Executive Summary
D1 Capital's Dan Sundheim manages over $30 billion across public and private markets, with significant positions in OpenAI, Anthropic, and SpaceX. His core thesis challenges conventional wisdom about hyperscaler durability. Sundheim argues that AWS, Azure, and GCP face structural disruption as AI workloads become the majority of cloud computing. The LLM providers—concentrated among 4-5 companies—will eventually generate enormous free cash flows and economically justify in-sourcing compute infrastructure. This represents a shift from fragmented enterprise customers to concentrated AI providers who are better at GPU cluster management than traditional hyperscalers. Sundheim draws parallels between LLM business models and Netflix/Spotify: massive upfront capital investment in fixed assets (models/content) amortized over growing user bases with high incremental margins. The key difference is personalization creating switching costs, similar to Spotify's data advantage despite commoditized music. He believes scaling laws will continue, making AI the ultimate productivity tool driving economic growth with disinflation. However, the capital intensity introduces unprecedented financial and operating leverage, requiring sustained adoption speed to justify returns. Sundheim's contrarian view extends to software disruption, where he expects the first wave of AI shorts after years of only AI longs, starting with coding productivity tools like Claude Code threatening traditional software margins.
Key Insights
what Dan Sundheim said“AWS, Azure, something said GCP, their customer base was like every corporation in the world. Therefore, they had fragmentation and they had the benefits, massive economies of scale that no single company could get, and it was a very good business. The problem going forward is that I think that economically, it's highly unlikely that LLMs are not very concentrated in the hands of four or five companies.”
what Dan Sundheim said“I said, look, I think your business is some kind of combination between Netflix and Spotify. Netflix in that, unlike other tech companies, you are spending a ton of money upfront to train these models. Once these models are trained, you go sell them at extremely high incremental margins... The differentiating factor will not necessarily be that Google gives you a better answer... However, the personalization matters.”
what Dan Sundheim said“There haven't been any shorts in AI. There was like basically no shorts prior to 2026, really... In our letter, I said like, there are going to be a lot of shorts, some longs because of AI. Software is the first one.”
This is a preview. Log in to see the full analysis including investment opportunities, risks, catalysts, and detailed insights.