Manfred Bremmer
Editorial Manager Computerwoche

AI spending to reach $1.5 trillion this year, Gartner says

News Analysis
Sep 17, 20254 mins

Once it gets rolling, the AI avalanche is unstoppable. In 2026, global AI spending is forecast to top $2 trillion, according to Gartner.

Lawine, Schnee, Berg, 16:9, slider
Credit: Marcus Placidus - shutterstock.com

Companies can no more sit out the AI hype than they could the internet or cloud, according to the latest from Gartner. The market research firm expects global spending on artificial intelligence (AI) to reach almost $1.5 trillion this year. By 2026, Gartner expects global AI spending to exceed $2 trillion, largely driven by the integration of AI into products such as smartphones and PCs as well as the underlying infrastructure.

“We expect investment in AI infrastructures to continue to expand. Large hyperscalers continue to invest heavily in data centers with AI-optimized hardware and GPUs to expand their services at scale,” explains John-David Lovelock, distinguished vice president analyst at Gartner. “The investment landscape in AI has also long gone beyond the traditional US technology giants. Chinese companies and new providers of AI cloud services are also increasingly entering the market. At the same time, the continued commitment of venture capitalists to AI companies is providing an additional boost to global AI spending.”

Spending on AI-optimized servers is increasing rapidly

The predicted growth in AI-optimized servers is particularly striking for Lovelock. Spending in this area is expected to more than double to almost $330 billion within two years.

“The level of spending on this is staggering when you consider that practically everything in the world today runs on traditional servers – be it the banking system, the Visa system for travel, or all the services operated by Amazon, Microsoft, Google, Tencent, Alibaba or Baidu,” explains Lovelock in an interview with Foundry’s Computerwoche. “And yet, by 2027/2028, we will have spent more money on AI-optimized servers than ever on traditional servers.”

The number of servers purchased by banks, retailers and hyperscalers will remain unchanged, adds Lovelock. “With generative AI, these special genAI servers basically bring in additional new business. They don’t cannibalize existing investments and they don’t displace other investments –it’s a completely new spend.”

Global AI spending by sector 2025-2026 (million dollars)

Area202420252026
AI services259,477282,556324,669
AI applications83,679172,029269,703
AI infrastructure software56,904126,177229,825
GenAI models5,71914,20025,766
AI-optimized servers (GPU and non-GPU AI accelerators)140,107267,534329,528
AI-optimized IaaS7,44718,32537,507
AI chips138,813209,192267,934
AI PCs (ARM and x86)51,02390,432144,413
GenAI smartphones244,735298,189393,297
Total AI expenditure987,9041,478,6342,022,642

Source: Gartner (September 2025)

AI features: often sold, not selected

In market segments such as AI PCs, on the other hand, it is very clear that AI is largely sold along with the product and not specifically selected, according to the Gartner analyst. “When people go into a store today to buy a new PC, they are generally not specifically looking for an AI-optimized or AI-capable device – they simply want the best model for their budget.”

It is now often the case that AI is already integrated into these devices, explains Lovelock. “The smartphones being sold today also have AI functions integrated – but users can’t really do anything with them at the moment.”

Companies cannot sit out AI

All in all, the Gartner analyst sees the figures as a clear sign that AI is inevitable. “We are no longer in a situation where people are debating whether AI will come or not,” he explains. By 2028, it will be practically impossible to buy a phone, PC, tablet or laptop without generative AI being integrated. “Generative AI will be built into the operating systems that control all our technology and will be included in every product available for purchase. The idea that you can wait or ignore the technology will then completely disappear.”

In conversations, Lovelock says he sometimes hears that people are not yet working on AI because there are more important issues at the moment. “Everything is important and needs to be dealt with. But the idea that generative AI can be put on the back burner –that doesn’t work.”

Lovelock warns that this uncertainty will not subside in the next few years: “The current level of uncertainty is exactly the level we will have to get used to over the next three years.”

This article originally appeared on Computerwoche.

Manfred Bremmer
Editorial Manager Computerwoche

Manfred Bremmer ist Editorial Manager der Marke COMPUTERWOCHE von Foundry in Deutschland. Nachdem er bereits während seines Studiums als freier Journalist für die CW tätig war, stieg er Mitte 2001 fest als Online-Redakteur ein und blieb. Im Laufe der Jahre war Manfred in verschiedenen Ressorts tätig, darunter Unternehmen & Märkte, Communications und Mobile & Apps. Neben seiner Tätigkeit als Editorial Manager befasst sich der studierte Ethnologe mit zahlreichen IT-Trendthemen rund um die Digitalisierung und betreut den TechTalk-Podcast.

More from this author