AI Tech Fuels Data Centre Boom, but Hyperscalers Face Risks

As AI transforms global technology needs, hyperscale data centres are expanding at a record pace.

However, according to a report from Moody’s Ratings, this growth brings mounting financial, technical and strategic challenges.

While demand for data-heavy applications like Gen AI fuels a surge in hyperscale infrastructure, the limits of power, location and capital investment now pose critical risks.

The challenge of sustainable expansion

Moody’s Ratings anticipates that data centre capacity could grow by 20% annually after 2028.

This would be a sharp jump from the moderate 5% or even fast 10% growth previously seen.

Hyperscalers – the large tech companies that build and lease massive server campuses – are at the heart of this expansion, funnelling investment into next-generation AI data centre campuses.

But as Moody’s notes, predicting when this growth will slow is becoming increasingly difficult.

“Surging growth in hyperscale data centre capacity will eventually level off, but identifying that inflection point has become increasingly difficult as AI data centre campuses emerge as another key growth driver,” the report states.

AI data centres now power much of this growth.

Unlike traditional facilities, AI data centres must be built with specialised infrastructure to meet the electricity and cooling demands of advanced computing.

Many of these new builds are located in remote regions, selected for reliable and lower-cost power access. Yet those same sites bring logistical problems – from labour shortages to high transport costs.

Furthermore, data centre infrastructure now needs to deal with unprecedented heat output.

Legacy facilities are not equipped to manage these thermal loads, which makes upgrades or replacements essential.

Rajesh Sennik, Head of Data Centre Advisory at KPMG UK, told our sister publication, Data Centre Magazine, at the end of 2024: “The industry is now facing unprecedented demand for new infrastructure solutions to efficiently power, cool and support this next generation of compute and as a result, AI is fundamentally reshaping the architecture of IT infrastructure.”

Gigawatt campuses and strategic uncertainty

Hyperscalers are constructing AI data centre campuses on a scale not seen before.

These sites centralise gigawatts of computing power into single developments. Projects such as the 5.6GW Wonder Valley in Alberta, Meta’s 2GW Louisiana site and OpenAI’s Stargate AI joint venture are expected to add multiple gigawatts of capacity by 2029.

But much of this investment is speculative.

Moody’s cautions that as the infrastructure becomes available, it may not align with actual computing demand.

“As investment continues to pour into new data centres, some level of capital reallocation and retrenchment is inevitable,” the report warns.

“The hyperscalers that have been spurring the market's expansion continuously right-size their newly leased and owned capacity under development because much of this new capacity is being built in anticipation of future needs.”

This misalignment could affect future utilisation rates, operational efficiency and financial returns.

Data centres built to support a single tenant’s needs may need expensive upgrades if that tenant leaves or changes computing demands. Matching facility design with AI’s evolving requirements is now a critical factor in maintaining long-term value.

Moody’s underlines the capital risk: the upfront cost to build data centres capable of supporting generative AI far exceeds previous tech waves.

With average rack densities climbing to 12kW in 2024, and some predicted to reach between 1MW and 5MW in the coming years, power efficiency and adaptability are more important than ever.

Supply chain volatility and rising costs

The cost to build and maintain AI data centres continues to rise.

This is not just due to technical demands. Trade policy and geopolitical tensions are also reshaping the supply chain, with knock-on effects for construction and operations.

US tariffs on imported steel and rare earth minerals are pushing up material costs.

Moody’s says these pressures may force developers to delay projects or re-evaluate capital plans. Tariffs on essential electrical components could create additional bottlenecks.

“The emergence of Chinese start-up DeepSeek, which startled the tech sector earlier this year by announcing that it had trained its competitive AI chatbot with fewer and less advanced chips than those used by American rivals, illustrates how innovative breakthroughs can put capital investments at risk,” the report notes.

These rapid breakthroughs in AI also create a moving target.

Infrastructure built today might be outdated within a few years, depending on whether newer models require different chipsets, cooling techniques or energy strategies.

For hyperscalers – including the likes of Google, Amazon and Microsoft – these risks do not necessarily slow investment. But they do introduce new pressures on margins and operational flexibility. Cash-rich companies can weather short-term shocks, but long-term cost forecasting is becoming more difficult.

Moody’s explains: “Abrupt changes in tariff policies increase uncertainty, making it challenging for data centre developers to forecast costs and timelines accurately.”

Balancing innovation and investment

While AI continues to push the boundaries of what technology can do, hyperscale data centre development must strike a delicate balance.

Building ahead of demand carries risks – both financial and structural – but failing to anticipate AI’s growth could leave providers lagging.

As the market evolves, hyperscalers must find ways to deliver infrastructure that is both future-proof and adaptable. This includes taking account of rising rack densities, energy constraints and geopolitical trade impacts.

AI is rewriting the rules of data centre architecture, but Moody’s concludes that the key question is not how fast the industry can build, but whether it is building the right infrastructure at the right time.

Explore the latest edition of AI Magazine and be part of the conversation at our global conference series, Tech & AI LIVE. 

Discover all our upcoming events and secure your tickets today. 

AI Magazine is a BizClik brand