
AI infrastructure spending is rapidly centralizing around frontier labs like OpenAI and Anthropic, which are shifting from venture-funded losses to generating massive revenue per megawatt. These labs now command an increasing share of global incremental compute, with projections suggesting they will control a majority of usable flops by 2028. This concentration is driven by economies of scale in training and the ability to outbid other market participants for scarce hardware. While the supply chain for EUV tools and power generation faces physical and capital constraints, the high return on compute investment incentivizes further expansion. As these models approach the capability of automated software engineers, the opportunity cost of capital rises, potentially triggering broader economic shifts, including higher interest rates and the crowding out of traditional industries. This trajectory points toward a future where a handful of entities control the vast majority of global AI labor and compute resources.
Sign in to continue reading, translating and more.
Open full episode in Podwise