Data centres: Powering the AI revolution through private capital
Atul Roy, Managing Director, Head of Telecoms Infrastructure at Cordiant Digital Infrastructure
The world is waking up to the transformative potential of artificial intelligence (AI). Adoption of the technology has been widespread, allowing firms of all sizes and industries to profit from efficiency gains and embrace innovation. From automating time-consuming admin tasks, such as data processing, to supporting more high-level tasks, such as content drafting, the technology has proven to be inherently versatile. As the technology continues to develop, creating new use cases, the demand for AI shows no sign of slowing.
To accommodate this, investment in digital infrastructure, namely data centres, will be crucial. This is where private capital has a clear opportunity to power the AI revolution. Goldman Sachs has forecast that AI will drive a 165% increase in data centre power demand by 2030[1], with it being widely acknowledged that global demand for data centres will quickly outstrip supply.
While there are various high-profile, large-scale investment funds targeting digital infrastructure, there is arguably a shortage of capital at the lower end of the market. Given the fast-paced nature of innovation in this space, there is an urgent need for more private capital to be deployed to help bridge this gap.
Shifting the dial through private capital
Private capital can help to de-risk public sector investment, as investors spend more time understanding the nuances of the business in more depth. As the capabilities of AI continue to develop and, as the technology becomes ever more ingrained in our daily lives, institutional investors are increasingly drawn towards data centres when looking to diversify their portfolios. For example, commercial real estate investors are now turning to basic infrastructure data centres in growing numbers in an effort to spread risk.
Although the technology is still in its infancy, a significant quantity of private capital has already been deployed to support it. This trend looks set to continue, as US$1trn in capital expenditure is expected in the coming years to fund the data centres, chips, energy and infrastructure that power AI technology[2]. While our industry can be proud of the progress made to date, we must not shy away from the reality: there remains a large gap to plug.
Back in January, OpenAI, Softbank and MGX unveiled The Stargate Project to much fanfare. Announced at the White House by President Trump, the firms vowed to direct $500bn into AI infrastructure in the US by 2029[3]. It cemented the US’ status as the world leader in AI investment, with Stanford University’s 2025 AI Index Report reporting that the US saw $109.1bn in private AI investment in 2024[4]. However, soon after the announcement was made, Elon Musk, a top adviser to President Trump at the time, cast doubt over the funding. Of the $100bn initial investment, only $52bn has been pledged to date. The whereabouts of the remaining $48bn remains unclear.
Looking to the future: The challenges for private capital
Although The Stargate Project has undoubtedly driven more interest in digital infrastructure as an asset class, there is a question of where exactly $100 billion could be deployed over a 12-month period to generate a strong return on investment. Naturally, there are limits on how much infrastructure can be built in that timeframe. Similarly, data centres typically require a host of permits which local authorities and building inspectors can be slow to grant.
The investment opportunity is also determined by the pre-existing availability of supporting infrastructure. For example, the International Energy Agency has forecast that the electricity consumption of data centres will more than double by 2030[5]. With energy infrastructure taking between five and 15 years to construct, national grids are left playing catch up to meet the surge in demand. The strain on the grid has been exacerbated by the growing demand for electricity more broadly, driven by the likes of electric vehicles and other innovations.
Private investors, therefore, cannot consider a data centre project in isolation. They must also assess the proximity of the site to energy infrastructure, the reliability of the electricity supply and any pre-existing commitments from local government to invest in the national grid. Failure to do so can limit the potential for returns.
An emerging AI arms race
As the supply of electricity becomes increasingly expensive in Europe[6], developing markets can offer a competitive advantage. Developing markets typically have lower barriers to entry for investors, in terms of electricity prices, the price and availability of land and the comparative ease of getting building permits.
The result is an AI arms race, where governments are scrambling to attract inward investment. The US leads the way in terms of private investment, aided by states that are easing regulation. While developing countries might offer some competitive advantages, investors can be deterred by the currency risk inherent in emerging economies.
The European Union is looking to capitalise on this and make itself a global hub for AI investment. Back in February, Commission President Ursula von der Leyen launched the InvestAI initiative, designed to mobilise €200 billion for AI investment[7]. The initiative forms part of the EU’s broader AI Continent Action Plan[8], one of the largest public-private partnerships in this space. It is set to bring a new lease of life to the European AI ecosystem and increase the stream of private capital investment.
Private capital has a clear responsibility
The growth in data centre demand in the last few years shows our industry’s adaptability, reflected in significant private capital mobilisation. But we must go further.
Private capital has the ability to shape the future of digital infrastructure, a role that involves looking beyond digital infrastructure in isolation and extends into closing the growing gap between the demand for AI data centres and the limited supply of energy.
There is an urgent need for targeted private capital to flow into mid-sized and emerging market data centre projects, especially those that can scale quickly, ease pressure on public infrastructure and adapt to local regulatory and energy constraints.
Institutional investors must stand at the ready to meet this demand.
[1] AI to drive 165% increase in data center power demand by 2030 | Goldman Sachs
[2] Gen AI: too much spend, too little benefit? | Goldman Sachs
[3] Stargate: Tech giants announce AI plan worth up to $500bn – BBC News
[4] Nvidia chief says UK lacks digital infrastructure as Keir Starmer pledges £1bn for AI
[5] IEA: Data center energy consumption set to double by 2030 to 945TWh – DCD
[6] Key drivers for electricity price changes in Europe
[7] https://ec.europa.eu/commission/presscorner/detail/en/ip_25_467
[8] https://commission.europa.eu/topics/eu-competitiveness/ai-continent_en