The Infrastructure Thesis

NVIDIA dominated the first phase of the AI boom. When investors thought about AI, they thought about GPUs. High-performance chips were the bottleneck. More models meant more compute, and more compute meant more NVIDIA. But AI spending doesn’t stop at chips. As GPU supply expands, the bottleneck shifts. And when the bottleneck shifts, capital follows. So what comes after NVIDIA? 1. Semiconductor Equipment Chip demand doesn’t just benefit designers. It benefits the companies that build the machines that manufacture chips. As AI models scale, advanced packaging and memory production become critical. This creates sustained demand for semiconductor equipment makers. 2. Memory and Advanced Packaging AI models rely heavily on high-bandwidth memory (HBM). As training clusters grow, memory becomes a constraint. Investors looking beyond GPU designers often watch memory suppliers and packaging specialists. 3. Data Center Infrastructure AI is not just a chip story. It’s a data center buildout stor...