Who benefits from the AI power bottleneck?

Power scarcity serves as a convenient, often self-serving narrative for various stakeholders in the AI and energy sectors. Hyperscalers, such as Microsoft and Amazon, benefit from emphasizing power constraints to incentivize infrastructure development while maintaining optionality in procurement. Meanwhile, hardware OEMs and engineering firms directly profit from the resulting surge in capital expenditure and newfound pricing power. Conversely, independent power producers and natural gas companies may downplay the urgency of new capacity to protect their margins in a high-price environment. Shanu Mathew, a senior vice president at Lazard, highlights how these conflicting economic incentives often drive public discourse more than objective technical reality. As the industry grapples with the massive capital requirements of AI, distinguishing between genuine physical bottlenecks and strategic posturing remains essential for evaluating the long-term viability of the current energy-intensive growth cycle.