What was once a futuristic vision is now embedded in daily life: artificial intelligence has become commonplace in under two decades. Despite its rapid advancement, the technology is still in its early stages. By mid-2025, ChatGPT alone had 800 million weekly active users, with daily adoption expected to surpass one billion before year’s end. This explosive growth is opening investment opportunities across the spectrum and challenging the notion that AI is only the domain of large-cap firms.
AI use takes two main forms: everyday interfaces such as ChatGPT and Google Gemini, and large-scale training of models that learn from massive datasets. Both require specialized, high-performance infrastructure capable of handling enormous volumes of data. That reliance is fueling investment not just in data centers, but also in the critical systems around them—power generation, transmission networks, and advanced cooling technologies.
At the heart of this transformation is the reliance on huge data centers with massive processing power capable of supporting large language models (LLMs).
How the power gets used
According to Goldman Sachs, approximately 60% of the current global capacity of data centers is utilized by hyperscale and enterprise data centres owned and operated by very large cloud/AI platforms.
AI runs on hyperscale infrastructure – requiring vast numbers of highly specialized computer chips, ultra-fast connections to move data between them, and robust systems to deliver steady power and cooling.
Goldman Sachs estimates that data center capacity is likely to shift from around 60% currently, closer to 70% in the hyperscale and enterprise category. Underlying that is a shift in demand for cloud computing and general business functions and towards AI, which is currently 14% and expected to rise to 27% by 2027.
Source: Goldman Sachs, “AI to Drive 165% Increase in Data Center Power Demand by 2030,” 2025
Those who assume data centers are only for large-cap investors would be mistaken: in 2020, San Francisco-based GI Partners invested in DR Fortress, Hawaii’s largest data center operator. Earlier this year, Berkshire Partners backed Fleet Data Centers with a campus-based commercial model, while Toronto-based Novacap invested in Hyscale Data Centers through its new digital infrastructure fund.
Mid-market investors are also challenging the notion that capital must be tied up for the long term. In October 2024, Cordiant Capital’s digital infrastructure fund acquired a 47.5% stake in Datacenter United, a platform of 13 facilities across Belgium. Less than a year later, in July 2025, it syndicated €20 million of that interest to another institutional fund, reducing its stake to 37.4% while retaining joint governance and operational oversight.
Powering up for AI
Private equity isn’t just investing in digital infrastructure—it’s also exploring broader opportunities tied to AI’s rise. With more advanced applications, the demand for power is set to climb exponentially.
Since ChatGPT was first introduced at the end of 2022, the company has already released multiple versions of its model, with its most recent, ChatGPT-5, introduced in August 2025. Even so, roughly 80% of ChatGPT’s users employ it in a simple way, typically asking basic one-off questions. “That 80% of usage alone consumes the same amount of power as Germany,” says Mark Rossano, the founder and CEO of C6 Capital Holdings. He notes that if the same number of recreational ChatGPT users were to shift to more advanced usage, the AI platform would become the third largest power consumer in the world, behind only the United States and China.
“What a lot of people don’t appreciate is that as these [systems] get more powerful, they throw off more heat and you need more and more power to keep them cold, because operational efficiency is between 30 degrees Celsius to 40 degrees Celsius,” says Rossano. “When you look at where the power is going, it’s exponential. And it’s only going to continue to grow, because computational power is doubling every four months and that’s doubling the power demand almost every 1.5 months,” he says.
As all of the major technology companies vie to become the leaders in AI and are stepping up their AI training efforts the amount of data processed by AI models is exploding and is only likely to continue rising.
“With the data centers, where you’re going to have the need for a tremendous amount of power, there are grid concerns.
There are governmental concerns about the amount of energy that can even be produced, and it’s a subset of the energy space that’s very active,” says Mitchell Moses, a partner at law firm Duane Morris and founding partner of the firm’s Fort Worth, Texas office, pointing to additional ways that AI has changed the way the investment industry looks at the energy sector.
“When you talk about our energy group at Duane Morris it looks a lot different than it did 20 years ago… For PE companies and PE investors, AI is a new point on their checklist or their investment memo. And when firms do their SWOT analysis, or whatever their boilerplate is for their investment committee, AI is on the list,” he says, noting that energy industry concerns now include everything from renewables to remediation, environmental and even AI-based modernization.
As AI fuels the need for ever-greater amounts of power it is also fueling private equity investors’ interest in the businesses working to meet that demand. “We’re seeing a huge investment from private equity into two sides of the power story. One is generation, and the other is transmission because we need wires to move the electrons,” says Rossano, noting that these are the two areas that have been getting the most activity due to the global shortage of power.
To put the shortage into perspective, he points to PJM Interconnection, a Regional Transmission Organization (RTO) that runs the United States’ largest multi-state electricity transmission grid and a wholesale electricity market that serves 65 million people in 13 states. In PJM’s July 2025 auction, the price of power for the 2026/2027 delivery year hit a record price of $329.17 per megawatt day, up from around $120 just a year ago.
“We don’t have the electrons to support what’s happening, which is why you’ve seen Google, Microsoft and Amazon aggressively going after and buying nuclear facilities and buying up the massive hydroelectric facilities,” says Rossano.
From Mega-Deals to Mid-Market Plays
Deals such as Google’s $3 billion 20-year power purchase agreement with Brookfield Renewable Partners for power through two of its hydro facilities in Pennsylvania, and KKR and PSP Investments’ acquisition of a minority interest in American Electric Power for $2.8 billion have dominated headlines, but power-related investment opportunities are plentiful within the middle market as well.
“There’s a significant amount of opportunity in the $1 billion to $30 million range because there’s a huge amount of generation capacity that is overlooked,” says Rossano, whose own firm is investing in hydroelectric dams that range between one megawatt and 29 megawatts –investments that still add value, but which are too small for the private equity industry’s largest players.
Partners Group’s $450 million investment in PowerTransitions shows how private equity is backing transitional generation assets. Audax Private Equity’s acquisition of Total PowerGen Solutions highlights demand for backup power infrastructure, while Renown Capital Partners’ $60.3 million investment in Utilidata shows mid-market interest in smart grid technology to strengthen resilience as AI drives energy demand.
Beyond Generation: What’s Next
While generation is the most attractive power investment opportunity at the moment, there are many other promising areas within the sector that private equity investors should be looking at as well.
Given the difficulty of building transmission lines, fiber optics is an attractive workaround that enables facilities to disseminate power in a more affordable way. Another area that Rosanno points to as a promising investment opportunity is facilities able to purify what is known as sour gas –gas streams that are contaminated with high levels of sulfur –primarily hydrogen sulfide and carbon dioxide. As the gas supply becomes increasingly sour and existing treatment facilities near capacity, building new facilities to handle the upcoming problem now is an attractive proposition.
Beyond driving the need for more energy, AI is also altering the way that the private equity industry looks at the needs of the overall sector. One less traditional area of the energy sector that Duane Morris’ Moses believes is an attractive area for PE investors is technology-based service companies.
“What we’re hearing from our clients is that they’re looking for service-based companies that can be modernized with AI and other forms of technology. I think there’s going to be a lot of activity in that space as PE’s get their arms around how to implement AI and other technologies into making energy businesses safer and more efficient,” says Moses, adding that there are potential efficiencies AI could introduce to the industry as well.
“We’re not seeing it happen yet, but we have a lot of operating clients where the right PE firm could come in and increase profitability really quickly with AI,” he says.
It is clear that AI is transforming the energy landscape at remarkable speed, and with investors increasingly receptive to the sector, mid-market private equity funds have a significant opportunity to step in and capture value.
Stay in touch with all of our latest updates and articles. Sign up now.