Hyperscalers, vendors funding trillion dollar AI spree, but users will have to pay up long term
- Reference: 1768573100
- News link: https://www.theregister.co.uk/2026/01/16/hyperscales_and_vendor_fund_trillion/
- Source link:
The global tech analyst company predicts a 44 percent increase from last year's AI spending figure $1.76 trillion. This was revised up from a [1]September estimate of $1.5 trillion .
We are getting to the point where we go from 'that was a great idea' to 'where's my revenue?'
That $2.52 trillion figure will nearly double again, to $4.7 trillion, by 2029, Gartner predicted. However, John-David Lovelock, Distinguished VP Analyst at Gartner, told The Register that cloud hyperscalers and software vendors were likely to pick up the bulk of the bill in the short term – although enterprise IT may have to pay in the long term.
Users were also likely to employ enterprise software vendors' AI agents for initiatives, rather than build their own AI platforms with third-party technologies.
Lovelock said: "Hyperscalers are buying billions of dollars' worth of servers and trying to build the foundation for the next super cycle of intelligence. There are consumers who are happily taking on these chat clients, creating wonderfully inventive cat videos and purchasing those devices, the mobile phones, the PCs, the tablets that are all now AI-enabled.
[2]
"There's the regular enterprise consumer or the enterprise buyer, whether that's the CIO or the board-appointed special group who are saying, 'Get me something AI'."
[3]
[4]
However, enterprise users are in the "trough of disillusionment" with AI, as project failure rates hitting about 90 percent, he said.
"We're starting to see the end of the investment line. We had a thousand flowers blooming, now it's time to prune the garden. We are getting to the point where we go from 'that was a great idea' to 'where's my revenue?' That's a normal part of any new technology," he said.
[5]
During the coming period, enterprise users are more likely to turn to their trusted software vendors to build AI into their solutions rather than attempt to stitch a solution together themselves with third-party software.
"They're going to be looking for solutions that come from their incumbent providers, both software and services," Lovelock said.
For example, Salesforce already had Einstein, ChatGPT and Agentforce built into its software portfolio, he said.
[6]
"It gives users the opportunity of a low-risk project if, in some cases, the software is coming as part of an update, they may not even be paying [extra] for it," he said.
In December last year, Salesforce's chief revenue officer Miguel Milano said he was relaxed about the CRM giant losing money on fixed-price AI agent contracts in the short term because it would be able to claw back the investment in the long term.
[7]Over half of AI projects are shelved due to complex infrastructure
[8]Contagious Claude Code bug Anthropic ignored promptly spreads to Cowork
[9]Developer writes script to throw AI out of Windows
[10]Block CISO: We red-teamed our own AI agent to run an infostealer on an employee laptop
"I have another 20 years to monetize that customer," [11]Milano said late last year.
Lovelock said it was fair to encapsulate SaaS software vendors' approach as being as happy to take short-term losses on AI investment because they would have plenty of opportunities to make the margins that their investors expect in the long term.
But he said vendors were also making a defensive play.
"If you're a Salesforce, you have to have AI in order to defend your CRM revenue. This year, more than half of the money being spent on enterprise application software will be spent on applications with GenAI in them. If a vendor's product does not have GenAI, the market expectations for growth are negative. It's a collapsing market without AI," Lovelock said. ®
Get our [12]Tech Resources
[1] https://www.theregister.com/2025/09/17/gartner_ai_spending/
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2aWpuvGUpTMwko5BdQgwEegAAAlc&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44aWpuvGUpTMwko5BdQgwEegAAAlc&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33aWpuvGUpTMwko5BdQgwEegAAAlc&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44aWpuvGUpTMwko5BdQgwEegAAAlc&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33aWpuvGUpTMwko5BdQgwEegAAAlc&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[7] https://www.theregister.com/2026/01/15/half_ai_projects_shelved/
[8] https://www.theregister.com/2026/01/15/anthropics_claude_bug_cowork/
[9] https://www.theregister.com/2026/01/13/script_removes_ai_from_windows/
[10] https://www.theregister.com/2026/01/12/block_ai_agent_goose/
[11] https://www.theregister.com/2025/12/15/salesforce_ai_monetization/
[12] https://whitepapers.theregister.com/
Who is getting rich?
So if trillions of dollars are being spent on AI who is actually earning that money? Is it the DC builders and hardware manufacturers?
Re: Who is getting rich?
Yep. Mostly it's Nvidia, sometimes with systems builders like Dell or Supermicro.
Some DC builders may do OK, however some are in a risky position where they keep ownership of the building and lease it back to the company that ordered it, such as Google, based on the "promise" of massive AI usage that will definitely come any day now. And by "builder" here I mean the prime contractor. I hope that the subs pouring concrete and laying cable are getting paid up front...
where we go from 'that was a great idea' to 'where's my revenue?
Thanks goodness for Gartner being around to provide statements of the bleedin' obvious, we'd never have figured it out otherwise...
How screwed are we in the event that nobody, except in a few rare cases can figure out how to sufficiently monetize non-free AI? Or for that matter AI that costs orders of magnitude more as AI companies try to claw back their investments.
wonderfully inventive cat videos
And that is worth a few trillion dollars, isn't it.
Nobody is going to pay for this. The 'investors' were conned. Their cash is nestling safely in the accounts of tech sector companies who will throw in the towel and move on, keeping the loot.
Best scam since the South Sea Bubble.
"I have another 20 years to monetise that customer," said the guys running Atari, Commodore and Psion.
To protect themselves from malware and geopolitics, and reduce their costs, may I suggest that the Next Big Thing will be standalone, offline machines with basic software packages, devoid of AI, SaaS and the cloud. Now that might be worth investing in, early stage,
"they may not even be paying [extra] for it,"
Oh yes they will.
"I have another 20 years to monetize that customer,"
Assume much?
Just a thought. We keep saying Europe needs its sovereign data centres and the nay-sayers saying it would be impossible to compete with the US incumbents. If they have tied themselves up with the expense of AI it should make it easier to compete simply by just keeping the operation clean. Even better if there's a fire sale of facilities to clear debts.