Exactly the same! Yet another "data center shutdown" story triggers a new round of corrections in U.S. AI stocks?

A piece of news about the suspension of a data center project is prompting the market to re-examine the pace at which US AI infrastructure deals are materializing. On June 10 local time, AI infrastructure company Crusoe announced it would suspend the construction of its large 1.8 GW data center project in Cheyenne, Wyoming (codenamed "Project Jade") at the request of an undisclosed client. Following the news, shares of Bloom Energy, which has supply chain ties to the project, plunged over 9% that day, closing at $235.92 with a daily drop of $23.69. The Nasdaq and S&P 500 indices also closed lower, with the technology and AI sectors under pressure overall. Goldman Sachs partner Rich Privorotsky immediately referenced the event in internal communications. He said the news “may have caught people’s attention” and directly pointed to a core market vulnerability: “In a market where everything is tied to AI capital expenditures, even isolated delays, postponements, or changes in priorities are enough to force investors to re-examine assumptions about future demand.” How large is this project? What does the suspension mean? What does 1.8 GW mean? 1 GW equals 1,000 megawatts, a capacity usually associated with a large power plant, not a single data center. Crusoe’s “Jade Project” was originally planned to include 900 MW of “behind-the-meter” energy in Cheyenne, specifically designed to bypass grid bottlenecks—showing from the start that the project made special arrangements for power supply and was not a typical engineering project. According to Data Center Dynamics, Crusoe halted construction at the client’s request. The identity of the client and the reason for the halt have not yet been disclosed. Rich Privorotsky also admitted, “The project has only just begun development and the client’s identity has not been disclosed, so it would be wrong to draw broad conclusions now.” But the issue is: The market isn’t waiting for conclusions. The stock price is already reacting. Why was Bloom Energy hit first? The reason Bloom Energy’s drop was so sharp is because its share price is built on a narrative highly dependent on demand realization from AI data centers. According to Investing.com, the “Jade Project” is linked to a conditional power purchase agreement with American Electric Power (AEP Energy), which is tied to Bloom’s solid oxide fuel cell supply. AEP disclosed in an SEC filing in January this year that its non-regulated subsidiary had agreed to purchase the majority of solid oxide fuel cell options for about $2.65 billion, and to sign a 20-year power purchase agreement with a high investment-grade customer to supply electricity from a planned fuel cell plant near Cheyenne. Any slack in this supply chain directly calls Bloom’s revenue outlook into question. Bloom’s fundamentals are not without support. In April, Bloom announced that Oracle planned to purchase up to 2.8 GW of fuel cell systems, with 1.2 GW already signed and deployed. Bloom’s revenue last quarter was $751.1 million, up 130.4% year-on-year, with product revenue jumping 208.4%. The company raised its full-year 2026 revenue target to $3.4–3.8 billion, a mid-point increase of about 80% from last year. But on Wednesday, these figures temporarily lost their persuasive power. Funds that had been speculating on the “S&P 500 inclusion” expectation the previous day quickly pulled out, and market attention shifted back to a more practical issue: When will orders actually convert into revenue? This scenario already played out once last year. This isn’t the first time AI infrastructure stocks have crashed due to a “construction site shutdown.” A similar market reaction happened to CoreWeave last year. According to The Wall Street Journal, CoreWeave’s market capitalization shrank by $33 billion in six weeks and its share price dropped 46%. One trigger was a delay in the construction of a large AI data center in Denton, Texas. Denton, north of Dallas, saw summer rains and strong winds delay construction by about 60 days, preventing the contractor from pouring concrete. That project was about 260 MW, which CoreWeave planned to lease to OpenAI. Completion was delayed by several months, and design adjustments caused further holdups. Communication issues worsened the stock pressure. CoreWeave CEO Michael Intrator said at the November 10 earnings call that the issue was limited to “one data center”; the CFO later corrected this to “one data center supplier.” Intrator repeated “one data center” in a CNBC interview the next day, correcting himself only after the host prompted him. CoreWeave’s stock fell 16.3% that day and continued downward in December. CoreWeave’s model makes investors more sensitive. It buys a large number of Nvidia’s advanced AI chips with high-yield debt, deploys them in data centers, and then rents out the compute power to customers, who are concentrated among a few big clients—OpenAI, Microsoft, and Meta. CoreWeave's sales last quarter rose from $583 million a year earlier to nearly $1.4 billion, but the company is still unprofitable, losing $110 million in the quarter. DA Davidson’s Gil Luria said CoreWeave’s balance sheet is “the ugliest in tech,” and that its operating margin is about 4%, less than half the cost of its debt interest. CoreWeave’s example shows that AI infrastructure deals require more than orders and client lists—they depend on construction, power supply, financing, and delivery. Any slowdown in any of these links can transmit directly to valuations. Goldman Sachs warning: Cyclical risk is accumulating Rich Privorotsky’s comments do more than explain Bloom’s drop; they point to a more systemic issue. Currently, market momentum returns are at the 90th percentile for the past five years, and overall exposure at the 99th percentile. As leveraged demand rises, financing spreads have widened, and retail participation in leveraged ETFs remains substantial. “All things are increasingly tied to AI spending. It’s part of the hardware long trade, it’s driving part of GDP, and it’s powering much of market performance. The cyclicality is becoming harder and harder to ignore.” His view: If anything interrupts the AI spending cycle, the vulnerability is obvious. This risk has existed for a while, but the market is accelerating toward it. That’s also why the Project Jade suspension is being amplified. The project itself had just started; the client hasn't been disclosed; but the market’s positioning and narrative have already been fully leveraged on the continuation of AI capex expansion. As soon as there are “delays”, “pauses”, or “priority changes,” investors will ask: Will future demand continue to accelerate, or will the pace of realization fall short of valuation assumptions? Privorotsky also breaks down the uncertainty of the AI investment cycle into two explanations. On the bullish side, the overall “pie” continues to get bigger, requiring more edge computing, more data centers, more memory, more electricity, and more networks. On the bearish side, many economically valuable tasks may ultimately—or already can—run on existing hardware, and the demand boom could come later than current valuations imply. He writes that the debate is less and less about model quality, and more about where inference ultimately takes place: in high-cost centralized cloud, or on cheaper, more open, and localized models. This is the core of the current adjustment. It’s not about a single project determining the AI cycle, but the market re-assessing whether AI capex, data center construction, power support, and corporate revenue can all keep in step as previously assumed. Meanwhile, Crusoe is trying to stabilize market sentiment. The company said Tuesday it had signed a total of 4.9 GW of AI infrastructure agreements, covering its data center projects and Crusoe Cloud business, with a total development pipeline exceeding 40 GW. CEO Chase Lochmiller said, “The demand from the world’s leading tech companies for AI infrastructure—fast and at scale—has never been stronger.” But given the suspension of the “Jade Project,” the market is currently skeptical. Risk Warnings and Disclaimer The market carries risks; investments should be made with caution. This article does not constitute personal investment advice, nor does it consider an individual user’s specific investment objectives, financial situation, or needs. Users should consider whether any opinions, views, or conclusions in this article are appropriate to their particular circumstances. Investing based on this article is at your own risk.