The rapid advancements in artificial intelligence (AI) are fueling an unprecedented demand for specialized computing components such as high-performance GPUs, CPUs, storage devices, and networking equipment. Industry players developing AI technologies are purchasing essential hardware like Nvidia GPUs and Micron memory chips in extraordinary volumes, triggering a notable increase in the stock valuations of companies specializing in AI hardware.
Nevertheless, hardware alone does not represent the entire landscape of investment opportunity within the AI ecosystem. According to a McKinsey estimate, total capital expenditures linked to AI infrastructure are projected to reach approximately $5.2 trillion through 2030. A significant portion of these allocations must be directed toward building and expanding the physical infrastructure that supports the deployment and operation of AI systems. This includes constructing data center shells, scaling existing power grids, and enhancing electricity generation capacities to meet escalating demand. Below is an exploration of some vital yet often overlooked infrastructure sectors instrumental to the AI expansion.
The tangible backbone: AI data centers
AI workloads necessitate distinctive computing environments often referred to as AI factories. These are sophisticated, purpose-built data centers designed to host critical hardware components for AI processing. The inherent complexities of these facilities extend beyond housing chips; they require advanced cooling technologies and access to substantial, reliable electrical power to sustain their operations efficiently.
Entities specializing in the development and management of such data centers are becoming increasingly integral to the AI value chain. For instance, data center Real Estate Investment Trusts (REITs) like Equinix and Digital Realty are central players in real estate and facility development tailored to AI needs. Equinix is actively expanding its portfolio of xScale data centers, which are large-scale hubs optimized for AI workloads. In 2024, it established a $15 billion joint venture focused on acquiring land and constructing multiple advanced xScale facilities.
Similarly, Digital Realty has initiated its first Hyperscale Data Center Fund in the United States in 2025 to underwrite up to $10 billion for data center projects. Furthermore, Digital Realty engaged in a partnership with Blackstone in late 2023, creating a $7 billion joint enterprise to develop large-scale data centers strategically positioned to meet AI growth demands.
Another significant stakeholder, Brookfield Infrastructure, has incrementally built a robust data center platform through acquisitions on a global scale. Presently, Brookfield operates over 140 data centers with a combined capacity of approximately 1.6 gigawatts (GW). The firm envisions expanding this capacity by an additional 3.4 GW across its portfolio. Beyond mere facility expansion, Brookfield is investing in complementary data infrastructure technologies, including plans to implement advanced fuel cell power sources developed by Bloom Energy to enhance data center energy efficiency and reliability.
Powering the AI era: electricity demand and supply challenges
AI’s computational demands translate into extraordinarily high energy consumption. The cutting-edge GPUs, intensive cooling systems, and associated data center technologies require vast quantities of electricity. AI data center campuses frequently demand power exceeding 1 gigawatt, an amount roughly equivalent to the electrical consumption of 750,000 average homes. Anthropic, a prominent AI company, estimates that U.S. AI power needs will reach at least 50 GW by 2028.
Meeting this sizable energy requirement calls for substantial investment in energy infrastructure. This encompasses expanding transmission lines, augmenting natural gas pipeline networks, and developing renewable energy generation assets. Numerous companies within the energy sector are actively scaling their capabilities to support this burgeoning demand.
NextEra Energy leads the push to enhance electricity transmission infrastructure, with plans to invest over $25 billion in forthcoming projects intended to reinforce grid reliability and capacity. The company is simultaneously expanding its natural gas pipeline system to meet increasing gas power consumption. Additionally, NextEra is at the forefront of renewable energy development. Notably, it has forged collaborations with Google aimed at accelerating nuclear energy deployment and co-creating extensive data center campuses.
Williams Companies, another key player, is advancing multiple projects to bolster natural gas supply capacity nationwide. These undertakings are projected to come online by 2030 to facilitate the rising energy needs linked to AI. Williams is also reviewing an extensive pipeline of potential projects amounting to over $14 billion earmarked for development between 2027 and 2033. Moreover, the company is actively constructing approximately $5.1 billion worth of gas-powered generation assets tailored for data center clients and maintains a portfolio of innovative power initiatives under consideration.
Concluding observations on AI infrastructure investing
Artificial intelligence stands to be one of the most transformative technologies in modern history, unlocking vast economic and societal benefits. Yet this potential cannot be fully realized without the necessary physical infrastructure that enables the effective deployment of AI hardware and software systems. From dedicated data center developers like Equinix, Digital Realty, and Brookfield Infrastructure to energy infrastructure leaders such as NextEra Energy and Williams, numerous companies occupy pivotal roles in building and sustaining the foundational structures underpinning the AI revolution.
Investors seeking exposure to AI’s expansion should consider these integral, though sometimes overlooked, sectors. The intersection of real estate development, energy provisioning, and advanced technological infrastructures presents an essential foundation that could support continued growth well into the future AI era.