Capital and Distribution in AI, and the Rise of Neoclouds with Kevin Zhang

Massive capital expenditure by hyperscalers in AI infrastructure hinges on future inference demand, creating a critical need for software revenue to eventually align with hardware outlays. Foundation models currently favor well-capitalized incumbents as training costs scale exponentially, though potential architectural shifts remain a wildcard for disruption. The US and Chinese tech ecosystems exhibit starkly different dynamics: the US market benefits from abundant capital and a risk-tolerant environment, whereas China’s capital scarcity necessitates a more cautious, "risk-off" approach that influences business models and monetization. Neoclouds like CoreWeave and Nebius are emerging as specialized GPU-as-a-service providers, using hardware access as a wedge to challenge traditional cloud giants. Investor Kevin Zhang emphasizes that while AI adoption will likely follow a five-to-ten-year trajectory, the long-term success of these ventures depends on balancing infrastructure costs with tangible enterprise ROI.