NEW YORK, July 8 — A new report from industry analysts at Gartner, published in March 2025, has laid bare a bottleneck in the race to build the artificial intelligence future: theft is stealing the very chips needed to power AI data centers, causing delays that could ripple through the global economy. The report estimates that 15% of U.S. data center construction projects in 2024 experienced delays due to theft, with the average delay clocking in at eight months.
The practical upshot: a project that was supposed to go live this year might not be ready until next year, and that has real consequences for the companies racing to deploy AI services. The total cost of these delays, according to Gartner, hit $3.2 billion in 2024.
That figure covers the replacement of stolen equipment and the extra labor needed to get projects back on track. The U.S. Census Bureau reported that data center construction spending reached $35 billion in 2024, so the theft-related costs represent a significant chunk of that investment.
The Chip Shortage Gets Worse
At the heart of the problem is the global AI chip shortage, which is already being driven by supply constraints on NVIDIA’s H100 processors. Those are the workhorses of the AI boom, the chips that train and run large language models like the ones powering chatbots and image generators. When thieves make off with racks of these GPUs — graphics processing units that can cost tens of thousands of dollars each — the replacement process is slow.
Gartner’s analysis notes that replacing stolen GPUs can take three to six months due to manufacturing lead times. In plain terms, if a construction site is hit today, the company might not see a replacement chip until months later.
That delay, the analysts predict, could push the completion of some AI data centers back by six to 12 months, further tightening an already strained supply chain. The practical upshot for ordinary readers is this: every delay in building a data center means a delay in the rollout of new AI features, from smarter virtual assistants to more capable image generators. It also means that the companies building these facilities are spending billions of dollars on security measures they hadn’t planned for.
Tech Giants Respond
Microsoft, Google, and Amazon have all announced increased security budgets for their construction sites in response to the thefts. Microsoft, in its 2024 annual report, allocated $500 million in 2025 specifically for site security. That’s the kind of spending that would have been unheard of just a few years ago, but it reflects the scale of the problem.
The U.S. Department of Energy has weighed in, noting that the delays could slow the deployment of AI infrastructure needed for national security applications. The implication is clear: these aren’t just corporate headaches; they have implications for government AI projects as well.
Gartner’s report also warns that the delays could reduce 2025 data center construction spending by 10% as companies reassess their security needs. That would be a significant pullback in an industry that has been growing at breakneck speed, and it could have knock-on effects for the broader economy. The bottom line?
The AI boom is running into a very old-fashioned problem: theft. And until the industry gets a handle on security, the chips that power the future will keep disappearing.
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “AI Data Center Thefts Delay Completion 6–12 Months, Gartner Warns”, “description”: “NEW YORK, July 8 — A new report from industry analysts at Gartner, published in March 2025, has laid bare a bottleneck in the race…”, “datePublished”: “2026-07-08T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Erika Johnson”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.infopulsetoday.com/”}}
