More than half of AI data centers currently under construction will fail in 2026. This is a worldwide phenomenon as the Gulf states are currently in crisis mode. AI data center output is tokens, and if there is a shortage of tokens, the cost to access AI will increase. The AI industry, from top to bottom, is on the verge of major restructuring as the ripple effect hits companies and consumers alike.
Ordinary consumers are already showing sensitivity to pricing signals. Google recently shifted from unlimited flat-rate AI access to a tiered, metered credit system — AI Pro at $19.99/month and AI Ultra at $249.99/month — signaling that even a company of its scale cannot sustain unlimited token consumption at flat rates. When prices rise meaningfully:
Casual users will churn first, abandoning AI tools they treat as novelties rather than necessities
Power users will consolidate to a single provider and become far more deliberate about usage, minimizing token consumption
Consumer AI features embedded in apps (writing assistants, image generators, chatbots) may disappear or get paywalled, reducing exposure without users actively canceling anything
The most dangerous scenario is a demand destruction spiral: higher token prices → reduced enterprise adoption → lower revenue for AI labs → less capital for data center expansion → even tighter compute supply → further price increases. Anthropic CEO Dario Amodei has warned that being “off by a couple of years” in infrastructure planning “can be devastating,” and that some firms haven’t thoroughly analyzed their financials.
The industry is walking a tightrope between subsidizing growth and reaching sustainable unit economics — and a construction slowdown would cut the rope. ⁃ Patrick Wood Editor.
Compass Datacenters has decided to withdraw from its plan to develop a major data-center corridor in Northern Virginia after spending years pursuing approvals and investing tens of millions of dollars, according to Bloomberg.
The company ultimately concluded the project wasn’t feasible due to mounting legal challenges, stricter regulations, and weakening political support, particularly around tax incentives.
This move highlights a broader shift in how communities and policymakers are responding to data-center projects. Local residents have increasingly raised concerns about issues like energy consumption, environmental impact, and potential effects on property values. As a result, companies in the industry are finding it more difficult, expensive, and time-consuming to gain approval for new developments.
Bloomberg writes that the proposed project was part of a larger effort to expand Northern Virginia’s role as a global hub for data centers. However, conflicts over land use, public notice procedures, and zoning approvals led to court rulings that invalidated key permissions. Faced with the prospect of prolonged legal battles and uncertain outcomes, Compass chose to step back.
The situation also reflects growing political sensitivity around how much support these developments should receive. Debates over tax breaks and incentives have made officials more cautious, while organized community opposition has become more influential in shaping decisions. Together, these pressures are forcing companies to rethink where and how they expand.
Meanwhile, another developer involved in the broader plan is still considering whether to continue challenging the rulings, showing that while some companies are retreating, others may continue pushing forward despite the growing resistance.
Recall days ago we wrote that half of US data centers scheduled for 2026 would be cancelled or delayed. We wrote then that the outlook for the US AI revolution looks increasingly more dim.
That’s because, as Canaccord Genuity analyst George Gianarikas writes, “the American data center boom is hitting a formidable wall of logistical friction.” He is referring to the latest outlook by Sightline Climate, which is also reinforced by recent articles from Bloomberg and others, and reveals a sobering reality for 2026: nearly half of the nation’s planned 16-gigawatt capacity faces cancellation or delay, with only 5 gigawatts currently under construction.
This inertia stems from a volatile mix of local permitting hurdles, community resistance, and a desperate reliance on overextended global supply chains for critical components like transformers and helium.
That’s right: half.
That’s right: despite $700BN+ of expected 2026 hyperscaler capex, nearly half of the data centers scheduled to begin operations in the US in 2026 “will either face delays or outright cancellations.”
The data, which comes from Sightline Climate’s 2026 Data Center Outlook, suggests that just 30% – 50% of the ~16 GW of planned US capacity for the year will face risks, with only ~5 GW currently under construction!
By 2027, the gap between ambition and reality widens further, as a mere fraction of the announced 21.5 gigawatts has actually broken ground. Worse, according to Futurism, data centers slated to open in 2027 are progressing far more slowly than anticipated. “Only about 6.3 gigawatts worth of computing infrastructure are actually under construction, compared to 21.5 announced gigawatts.”
And then visibility drops to virtually nothing beyond 2028 as uncertainty increases materially in the outer years. According to the article, “things get even dodgier in the coming years, with the vast majority of data centers planned for launch between 2028 and 2032 having yet to even break ground. There are a further 37 gigawatts of planned infrastructure which haven’t even received a firm completion date, only 4.5 [gigawatts] of which have actually begun work.”
This trend suggests an increasingly uncertain future for the industry, where power constraints and grid instability cast long shadows over projects slated through 2032.
Patrick Wood is a leading and critical expert on Sustainable Development, Green Economy, Agenda 21, 2030 Agenda and historic Technocracy. He is the author of Technocracy Rising: The Trojan Horse of Global Transformation (2015) and co-author of Trilaterals Over Washington, Volumes I and II (1978-1980) with the late Antony C. Sutton.
