The AI Datacentre Boom: $3 Trillion Spree – Bubble or Sustainable Growth?
Share
The Explosive Growth of AI Datacentres: A $3 Trillion Investment
The artificial intelligence (AI) revolution is reshaping the global economy, and at its heart lies a massive surge in demand for datacentres. A projected $3 trillion (£2.3tn) is expected to be spent on these vital infrastructure hubs, which serve as the central nervous system for AI tools like OpenAI’s and Google’s Veo 3. But is this unprecedented investment a sign of sustainable growth, or a bubble waiting to burst?
Despite concerns, the market currently displays few signs of imminent collapse. Nvidia recently became the world’s first $5 trillion company, while Microsoft and Apple’s valuations have soared. OpenAI’s valuation continues to climb, potentially leading to a $1 trillion flotation next year. Alphabet, Google’s parent company, reported record revenues, fueled by growing demand for its AI infrastructure.
The Rise of AI Datacentres: A New Industrial Revolution
Just as coal and steel fueled the Industrial Revolution, datacentres are now driving a new era of economic transformation. Cities like Newport, Wales, are hoping to capitalize on this trend. Microsoft is building a significant datacentre at Imperial Park, near Newport, offering a chance to tap into the economy of the future. Dimitri Batrouni, the Labour leader of Newport city council, highlights the importance of embracing this future rather than clinging to the past.
The Scale of Investment: A Spending Spree Across Tech Giants
The investment isn't limited to a few companies. Amazon, Meta (Facebook’s parent), and Microsoft are significantly increasing their spending on AI, projected to exceed $750 billion over the next two years. This includes not only staff costs but also capital expenditures like datacentres, chips, and servers. This spending spree, described by Manning & Napier as “nothing short of incredible,” is driving massive construction projects, such as the £4 billion Equinix centre planned in Hertfordshire.
Funding the Boom: Private Credit and Potential Risks
While the US tech companies (hyperscalers) are expected to cover $1.4 trillion of the $3 trillion investment through their own cashflow, the remaining $1.5 trillion needs to be sourced elsewhere. A growing reliance on private credit – a segment of the shadow banking sector – is raising concerns at the Bank of England and other financial institutions. Meta has already tapped into this market for $29 billion in financing for a datacentre expansion in Louisiana.
Gil Luria, head of technology research at DA Davidson, distinguishes between “healthy” hyperscaler investment and “speculative assets without their own customers.” He warns that the debt being used could pose structural risks to the global economy if it sours, particularly given the rapid depreciation of these assets.
Sustainability Concerns and Revenue Expectations
Analysts at Morgan Stanley predict global spending on datacentres will reach nearly $3 trillion by 2028, with revenues from generative AI (chatbots, AI agents, image generators) expected to grow from $45 billion last year to $1 trillion by 2028. However, doubts linger about whether businesses, the public sector, and individuals will generate enough demand – and be willing to pay for it – to justify these revenue expectations. A recent MIT study revealed that 95% of organizations are getting zero return from their generative AI pilots.
The Uptime Institute's Perspective: Hype vs. Reality
The Uptime Institute, which inspects and rates datacentres, suggests that many announced projects will never be built or will be populated only partially. Andy Lawrence, the executive director of research at Uptime, describes many announcements as part of the “hype machine.” He emphasizes that a significant portion of datacentres will be specifically intended to support AI workloads, or will mainly do so.
Beyond AI: The Broader Role of Datacentres
Microsoft clarifies that its Newport datacentre will serve purposes beyond AI. These datacentres also underpin everyday IT operations and provide “cloud” services, where companies rent servers instead of purchasing their own. They are the backbone of the digital world, supporting everything from online banking to streaming services.
Further Reading
For more information on the AI boom and its implications, explore these resources:
