IDC Makes Ebullient AI Spending Forecast Out To 2029

Ever since Nvidia reported its most recent financial results, where company co-founder and chief executive officer Jensen Huang said that there would be somewhere between $3 trillion and $4 trillion in spending on AI between now and the end of the decade, we have been on the prowl for any market research that backs up this claim or is its source.

We poked around a bit and found a forecast that IDC put out just before the Labor Day holiday, which is ironic given how much the future of AI is agentic and how this is largely going to have systems talking to systems and making decisions without humans in the loop except for oversight.

IDC is not saying much publicly about the forecast, of course, because it wants to charge $7,500 a pop for the report it has put together. But its market researchers did say a few things in its public statements promoting that report, which we can glean for some insight and used to compare with other forecasts. (We expect all forecasts to converge, more or less, as the enthusiasm for increased spending rises.)

First of all, IDC expects for the infrastructure buildout to continue through 2029, and that service providers – which includes what we call hyperscalers and cloud builders as well as tier two service providers and the neoclouds, we presume – will account for 80 percent of the infrastructure spending as they build out agentic AI platforms.

Some of this spending will be for AI-enabled applications as well as for brand new agentic AI platforms that in turn become the means of creating new applications. These are not necessarily the same thing. Infusing AI into Salesforce, SAP, or Oracle ERP suites is not the same thing as an upstart or an incumbent application supplier wiping the slate clean and building an application stack that is solely and completely AI agents. This is, according to many, the future of application software. We expect a lot of upheaval here, more than we have seen since the advent of the original ERP stacks and the unified approach SAP took 30 years ago that transformed the way enterprises deployed applications.

Add it all up, and IDC says that AI spending, driven by agentic AI, will hit $1.3 trillion in 2029 and will have a compound annual growth rate of 31.9 percent between 2025 and 2029, inclusive. If you assume smooth growth every year – which is likely not a perfectly valid assumption because markets rarely grow smoothly, particularly because the global economy changes – then AI spending will look like the chart above and as the number show in the related table.

What we have always contended is that recessions don’t cause technology transitions but they certainly do accelerate them. The funny bit is that if agentic AI is successful, its very deployment may create the recession that further accelerates the use of agentic AI as tools are used to replace human workers.

This could all happen very fast, or if the return on investment for chatty AI is not there, as some recent reports suggest, it may take longer as agentic AI matures and becomes the real deployment model that utterly and completely transforms the IT sector and therefore every business on Earth.

In this environment, as was the case with the Dot Com Boom, you don’t want to be the first big spender, but you sure as hell don’t want to be the last. As usual, timing is everything.