Simon Gordon, Head of Sales UK and Channel Islands – Institutional Capital Services, JTC Group, writes that in many ways the First FT newsletter of 23rd July, said it all: “Google’s AI spending has rapidly transformed the group from an asset-light business into a capital-intensive one.
“The company said free cash flow for the three-month period to the end of June [2026] turned to minus USD5.9 billion as it raised its spending forecast for data centres and other AI hardware.”
Google is not the only member of the AI front-runners working to transform intelligence dream into physical, highly profitable reality; capital expenditure is rapidly on the rise throughout the sector. In short, AI is no longer solely a story about models, applications, and productivity. It is increasingly about physical infrastructure.
For some years now, as the experience of the funds administered by JTC Group testifies, infrastructure has been one of the most reliable investment sectors, with the FTSE Global Infrastructure Index, for example, recording compound annual growth of 8.6 per cent in the five years to June 2026 (despite a negative start to the run in 2022 of -6.2 per cent). And it is reasonable to predict that the growth in AI spend will increasingly be directed towards infrastructure generating a significant number of private capital structures aiming to finance this second major growth spurt.
In short, despite currently volatility in the AI investment space where, in 2026 thus far, the NQINTEL (the Nasdaq CTA Artificial Intelligence Index) gained towards 53 per cent to 2nd June before falling back by 12.6 per cent to 22nd July, there is no let-up in demand for data centres and these are not the only infrastructure assets being driven ahead by the AI boom, given that to operate, data centres require power, grid connections, and a substantial water supply (cooling), not to mention transport links (especially important during the construction phase), and environmental services.
Estimates for AI-related investment now run into trillions of dollars, and just like Google (above), many companies are investing today for scale and market share with the expectation that cash flow, indeed profits will eventually follow. Indeed, profits have already materialized for a series of JTC clients who are already investing data centres.
PWC, which describes infrastructure investment as ‘the platform for accelerating human progress’, clearly has its eyes set on the big numbers and, via figures modelled by Oxford Economics, has forecast that global infrastructure spend will total above USD150 trillion (in 2023 dollars) by 2050 (much of which to be focused on Asian bridges and roads, and separately, rail). Interestingly, however, this analysis anticipates that annual investment in datacentres will level off, even decline, from an estimated USD250 billion next year, down to around USD100 billion by 2031, and USD50 billion five years later, before recovering to over USD100 billion pa from about 2040 (presumably as the first builds begin to become inefficient).
As stated, a part, probably a substantial part, of the financing for this AI infrastructure expansion will be funded via private capital funds of one form or another, but although new projects are successfully raising capital, it is clear that investor enthusiasm, at least in Europe, is lower than it was two or three years ago. JTC’s experience suggests that capital raised per fund per is down, and that furthermore, the time taken to launch a project is extending. These issues are particularly significant when associated with fund launches from new management teams rather than established names.
Various theories have been put forward to explain the current, slower market in Europe, which seems to be flying in the face of continuing AI hyperbole (and capex). First, there appears to be a significant geographical divide between the USA, and everywhere else. Recent calculations have, for example, suggested that US data centre capacity more than tripled in the last 12 months, with America now home to 93 per cent of global AI compute capacity. Moreover, returning to Google’s Q2 2026 figures, we find that Google Cloud revenue surged 82 per cent year-over-year, while its backlog nearly quadrupled during the same period, meaning demand already exceeds available capacity.
(It’s also noteworthy, according to the economists at Payden & Rygel, that despite tariffs, almost all of US high-tech equipment is imported, driving demand for AI hardware across Asia, while foreign customers pay American hyperscalers to run Al workloads on US infrastructure.)
More general factors include the high energy prices driven not only by the failure to guarantee safe passage through the Strait of Hormuz, but also by the increasing involvement of Houthi rebels who pushed oil prices back over USD100 on 24th July with attacks of two tankers off Yemen. Clearly these issues are not specifically impacting infrastructure investment, but with some exceptions they are increasing volatility, and therefore risk in the investment market as a whole.
That said, it is clear the future of AI will be nothing without massive infrastructure expenditure and therefore this sector of the investment market is more than likely to accelerate in importance. And this despite Eminem’s wise words: “The truth is you don’t know what’s going to happen. Life is a crazy ride, and nothing is guaranteed.”
