AI is projected to need a whopping 121 GW of U.S data-center IT power by 2030….what becomes the bottleneck first?

While AI might be able to scale at software speed, its required power, chips and raw materials cannot. So with U.S. data-center IT power demand projected to reach 121 GW by 2030, which physical constraint do you think could most hinder progress? Electricity? Builders? Copper? Silver? Something else? And do you think those bottlenecks present a potential investment oppo? Source: Data from McKinsey, July 31, 2026. Referenced in the iShares 2026 Thematic Mid-Year Update submitted by

CryptoAlerta — análise de criptomoedas e mercado em tempo real