Gartner now expects worldwide IT spending of about $6.39 trillion in 2026, up 14.5%, and $7.35 trillion in 2027, up 15%. That would be the first year above $7 trillion. When this page first ran in April 2026, Gartner’s number for 2026 was $6.31 trillion. It has raised the forecast every quarter since, and almost all of the extra money is going into AI data centers.
How the 2026 Forecast Kept Rising
| Gartner release | 2026 IT spending | Growth |
|---|---|---|
| Earlier forecast | $6.15 trillion | 10.8% |
| April 22, 2026 | $6.31 trillion | 13.5% |
| July 27, 2026 | $6.37 trillion | 14.2% |
| October 21, 2026 | $6.39 trillion | 14.5% |
Spending by Segment, 2026 and 2027
| Segment | 2026, $B | 2026 growth | 2027, $B | 2027 growth |
|---|---|---|---|---|
| Data center systems | 879 | 73.8% | 1,366 | 55.5% |
| Devices | 847 | 7.2% | 916 | 8.1% |
| Software | 1,464 | 15.2% | 1,679 | 14.6% |
| IT services | 1,557 | 4.4% | 1,633 | 4.9% |
| IaaS | 292 | 31.3% | 367 | 25.7% |
| Communications services | 1,352 | 4.1% | 1,391 | 2.9% |
| Total | 6,391 | 14.5% | 7,352 | 15.0% |

What Stands Out
- Data centers take about half the growth. Data center systems add nearly $490 billion in 2027, about half of the $961 billion rise in total IT spending. In April, Gartner expected this segment to reach about $788 billion in 2026. It now says $879 billion.
- Software stays strong. Software grows about 15% a year, and AI software is the fastest part of it. Gartner says AI infrastructure and AI software will shape the next $1 trillion of IT spending growth.
- Rented compute keeps booming. IaaS grows 31% in 2026 and about 26% in 2027, driven by the hyperscalers and the newer GPU clouds.
- Services and telecom barely move. IT services grows under 5% and communications services under 3%. Most of the new money goes elsewhere.
- Memory inflation. Gartner calls it “memflation”. Higher memory prices lift device and server spending in dollars, but squeeze affordability, and Gartner expects the effect to last through 2027.
What It Means for Investors and Buyers
The forecast confirms where the money goes: GPUs, servers, networking, memory and the power to run them. That is the demand behind Nvidia, the memory makers and the chip equipment suppliers. See why memory prices will not come down and the semiconductor equipment outlook. For the companies doing the spending, our monthly AI capex tracker follows hyperscaler budgets.
For enterprise buyers, the message is less comfortable. Device and server prices are going up because of memory costs, and services firms are under pressure as AI does more of their work. That squeeze is covered in Palantir and the Big Five of consulting. The risk to the forecast is the same as the risk to the AI trade: if AI revenue does not keep up with AI spending, data center budgets are the first to be cut. For more on technology markets, visit technologies.org.