Spending growth is slowing
Investment growth is forecast to fall from nearly 100% in 2026 to 54% in 2027 and 12% in 2028. Yet even as growth decelerates, the projected outlay keeps rising—and current AI revenue is not enough to cover the returns the investment requires.
Cloud revenue growth accelerated from 25% in 2024 to 48% in the second quarter of 2026. That is a meaningful shift, but it does not settle the larger question: whether AI labs such as OpenAI and Anthropic are growing revenue fast enough to justify the build-out. Expectations for those companies underpin much of the infrastructure construction and the financial instruments supporting it.
The build-out needs more than operating cash
Goldman estimates that spending already exceeds the money companies generate from current operations. That points to greater reliance on borrowing. Further investment could also be constrained by limited access to electricity, labor shortages and a shortage of memory chips.
The bank warned in June that consensus forecasts for spending were far too low. Its new estimate raises the stakes, but a larger infrastructure bill is not evidence of a larger business. I think the key uncertainty is whether revenue at AI labs can support the expectations and financing built around them before the physical limits on expansion start to matter.
Daily AI news
Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.
Only what matters — every day
Follow on X