BlackRock: AI's power problem
BlackRock CEO Larry Fink says the next bottleneck for AI isn't chips but electricity — global data‑centre power demand could surge ~165% by 2030 as firms build AI infrastructure. BlackRock execs also argue AI (not new tokens) will be the main institutional driver in crypto's next phase. (techi.com) (coinpaprika.com)
Larry Fink released BlackRock’s 2026 Annual Chairman’s Letter, titled “Growing with your country,” on March 23, 2026, while the firm reported about $14 trillion in assets under management. (blackrock.com) Consulting firms have put a dollar figure on the underlying buildout BlackRock referenced: McKinsey’s models estimate roughly $6.7 trillion of data‑centre capital expenditure by 2030 under a central scenario, with about $5.2 trillion of that tied to AI‑specific facilities. (datacentremagazine.com) McKinsey’s breakdown allocates material shares of that capex to IT equipment, site development and—notably—energy delivery, estimating roughly 25% of AI‑related capex would flow to power generation, transmission and cooling systems in some scenarios. (datacenterdynamics.com) Goldman Sachs research that feeds the policy and investment debate projects near‑term jumps in data‑centre electricity consumption—including a roughly 50% rise in demand by 2027 in its baseline scenarios—shaping utility procurement and grid‑planning assumptions. (goldmansachs.com) Industry moves reflect that capital view: a BlackRock‑led consortium including Microsoft and Nvidia agreed in 2025 to acquire Aligned Data Centers for about $40 billion to secure hyperscale, AI‑ready capacity and land/power positions. (datacenterfrontier.com) BlackRock has also been reported to be in active discussions with governments about financing upgrades to power and transmission capacity to support AI deployments, signaling coordination between asset managers and public planners. (tech.yahoo.com) At the same time, BlackRock’s head of digital assets, Robbie Mitchnick, told the Digital Asset Summit in New York that institutional clients are narrowing allocations to major networks such as Bitcoin and Ethereum while seeing AI‑related infrastructure and use‑cases as the most interesting areas for crypto integration. (coindesk.com)