OpenAI’s ChatGPT may have captivated the world with its launch in late 2022, but it’s 2024 that’s shaping up to be the real game changer for artificial intelligence as an investment opportunity. Suddenly, the term “AI infrastructure” has become the talk of the town, with managers and investors scrambling to unveil ambitious plans centered around this burgeoning asset class. When they say “AI infrastructure,” they’re primarily zeroing in on data centers—the lifeblood of AI operations. However, it’s becoming increasingly clear that this isn’t just a one-trick pony; investments in fiber networks and telecommunications towers are also poised to ride this wave.
What’s particularly fascinating is how the notorious energy demands of data centers have been flipped from a liability into a golden opportunity. Infrastructure managers, seasoned in the realm of clean energy development, find themselves uniquely positioned to tackle this challenge. This intersection of AI and energy could very well give rise to a new breed of large-scale investment funds. Take Global Infrastructure Partners, for instance; they recently announced plans to raise a staggering $30 billion dedicated to AI-compatible data centers and the necessary energy infrastructure. This isn’t just a flash in the pan; Brookfield Asset Management’s president, Connor Teskey, has also indicated that an AI infrastructure strategy is “near the top of our list… as we begin to think of new products and product development initiatives at scale.”
The sector is already witnessing jaw-dropping transactions, signaling its meteoric rise. A prime example is the recent sale of AirTrunk, a pan-Asia data center business, which was purchased by a consortium led by Macquarie Asset Management for around A$3 billion in 2020 and sold just four years later for a staggering A$24 billion to a Blackstone/CPP Investments partnership. Such profitable exits not only validate the investment thesis but also provide firms like Macquarie with the means to realize early performance fees from their funds.
Looking ahead, the numbers are eye-popping. Data centers and their associated power infrastructure are projected to require a whopping $2 trillion over the next five years, with half of that figure earmarked for the United States alone. This landscape is ripe for growth, especially in the higher-returning infrastructure equity space. Yet, as industry leaders like DigitalBridge CEO Marc Ganzi and KKR’s Raj Agrawal have pointed out, there’s also an increasing demand for lower-risk capital pools. Udhay Mathialagan, CEO of Brookfield’s Global Data Centre Group, noted that many data center firms are exploring capital recycling opportunities, indicating a shift toward stabilizing assets that could align well with traditional real estate investments.
Real estate managers are certainly taking notice. Legal firm Linklaters has identified over 50 real estate funds in the market targeting more than $50 billion, with roughly 40% of that amount—just over $20 billion—specifically aimed at data centers. The trend is clear: data center platforms are adept at attracting significant capital, whether directly or through co-investments, which could potentially complicate fundraising for traditional blind pool funds.
However, let’s not forget that amid all this excitement, there’s a fair share of hype. UBS has cautioned that “hot sectors need to be handled with care.” Their 2025 outlook emphasizes that while large hyperscale platforms in Tier 1 markets are likely to be the biggest beneficiaries of AI, investors should be wary of paying inflated valuations. Smaller data centers may offer more value, but local market dynamics, energy supply, and customer quality are crucial factors that can’t be overlooked.
Whether one views the current enthusiasm as justified or overblown, it’s clear that AI infrastructure is set to remain a hot topic well into 2025 and beyond. The investment landscape is evolving, and those who navigate it wisely stand to gain significantly.
