BlackRock Seeks $30B for AI Infrastructure, Eyes Compute Futures

At the Milken Institute Global Conference on Tuesday, BlackRock CEO Larry Fink argued computing power should be treated as a tradable asset and unveiled a $30 billion AI infrastructure initiative with partners including Microsoft, NVIDIA, xAI and MGX. He framed the issue as a shortfall of processing capacity rather than an investment bubble and proposed market structures that would let firms buy and sell future computing capacity.

Fink told attendees the main constraint for AI growth is processing power, not overvaluation. He compared computing capacity to other basic inputs and noted large AI workloads require continuous, measurable compute. Demand for processors, high-bandwidth memory and data-center space is growing faster than supply.

To address the gap, he proposed futures contracts for computing power that would allow suppliers to lock in revenue and businesses to secure capacity and hedge costs. Under the proposal, customers could lock future costs and guaranteed availability while data-center operators and chip suppliers would gain more predictable cash flows.

BlackRock’s infrastructure initiative has an initial funding target of $30 billion, company executives estimate, and could expand to about $100 billion with debt financing. Technology partners named include Microsoft, NVIDIA and xAI. BlackRock’s infrastructure arm supported a $10.7 billion contract with AES to secure power for AI operations, and a consortium led by the firm has advanced a roughly $40 billion bid for Aligned Data Centers.

The partnership will involve energy companies such as GE Vernova and NextEra Energy to plan for increased power generation and grid upgrades. Executives noted that expanding compute capacity requires parallel investments in electricity supply and distribution to keep data centers operating reliably.

Bruce Flatt, chief executive of Brookfield Corp., spoke on the same panel and predicted that an expansion of data centers, cloud services and AI will reshape economic activity over the next decade. Jensen Huang, founder and CEO of NVIDIA, commented that larger AI infrastructure will convert large data sets into actionable results and accelerate innovation. Sheikh Tahnoon bin Zayed Al Nahyan, chairman of MGX, said the partnership aims to accelerate innovation and productivity gains.

Industry data show growth in AI workloads is increasing demand for specialized chips, high-bandwidth memory and large-scale data centers. Providers and investors are securing land, equipment and power contracts to host clusters for training and inference workloads. The proposals outlined at Milken aim to create contractual tools and pooled capital to match that demand.