AI capital expenditure forecast to exceed the cost of building railways in both the U.S. and the U.K. — with the internet added on top

The headline reflects a striking projection from a major accounting firm: capital expenditure on artificial intelligence is now expected to outstrip the cumulative cost of building the railway networks in the United States and United Kingdom—and, when added together, even exceed the combined investment in those railways plus the buildout of the commercial internet. The scale of spending is not merely large; it is accelerating in a way that has no direct historical parallel.

What sets this AI buildout apart, according to the firm, is not the technology itself but the pace and concentration of investment. Railway and internet eras unfolded over decades, with costs distributed across many private and public actors. AI infrastructure, by contrast, is being deployed in a compressed timeframe, led by a handful of hyperscale companies. That creates a distinct economic signature: rapid capacity expansion before clear productivity gains are fully measured.

For markets, the relevance lies in cash flow allocation and follow-on demand. Investors will watch whether enterprise adoption and efficiency gains keep pace with the enormous upfront spending. Persistent gaps could raise questions about returns, while healthy uptake would validate the spending trajectory. The comparison to railways and the internet is less about certainty and more about magnitude—this is a capital commitment that could reshape balance sheets for years.