The bill for the Iran war—will the "AI bull market" pay for it?

The bill for the Iran war—will the "AI bull market" pay for it?

The Middle East conflict is creating a structural shock to the global AI industry, pushing already high-tech assets to the brink of systemic risk.

Energy sector shocks have been substantially transmitted to AI infrastructure. Shipping disruptions in the Strait of Hormuz have led to a halt in nearly one-third of global crude oil exports and one-fifth of natural gas exports. Brent crude oil prices rose 40% in a single month, and liquefied natural gas prices surged simultaneously in both European and Asian markets, with helium spot prices doubling. Key segments such as chip manufacturing and data center operations are facing cost re-evaluations.

Meanwhile, the rift between the AI investment boom and the fragility of the macroeconomy is widening. According to The Atlantic, by the end of 2025, virtually all of the U.S.'s economic growth will be driven by AI investment, but this single-point growth engine is facing multifaceted risk failures simultaneously. This year, core tech stocks like Google, Meta, Microsoft, Amazon, Nvidia, and Oracle have slumped by 8% to 27%, dragging down the overall stock market performance.

Investor Paul Kedrosky pointed out that the fundamental difference between the current AI crisis and the 2008 commercial real estate risk is that "all vulnerable nodes are intertwined." Coupled with energy shocks, debt accumulation, and supply chain disruptions, the AI industry is facing its first systemic stress test since it became an economic pillar.

Highly Concentrated Supply Chain, Energy Shock Hits the Core

The globally concentrated supply chain layout of the AI industry is now exposed to systemic impacts due to the blockade of the Strait of Hormuz.

Currently, the most advanced storage and training chips are supplied mainly by three Asian companies, and these economies rely heavily on imports of crude oil and large quantities of liquefied natural gas from the Persian Gulf. Essential materials for chip manufacturing, such as helium, sulfur, and bromine, are also highly dependent on imports from this region. The actual closure of the Strait of Hormuz has triggered multiple transmission mechanisms:

  • Critical material shortages: Tight helium supply is beginning to threaten the stable production of AI chips, with significant upward pressure on prices;
  • Rising operational costs: Higher energy prices are putting even greater financial pressure on large data centers that already struggle to make a profit;
  • Capacity expansion hindered: New data center projects face stagnation risk due to dual pressures of cost and supply chain.

Sam Winter-Levy, a technology and national security researcher at Carnegie Endowment for International Peace, pointed out that the Strait of Hormuz “is crucial to almost every aspect of the global economy, and the AI supply chain is not immune.”

Meanwhile, the conflict is eroding the foundations supporting the AI industry from both physical security and capital supply. As the Middle East conflict continues, Amazon’s data centers in the UAE and Bahrain have suffered attacks, and security conditions for Gulf countries—key nodes in the U.S. AI strategy—have sharply deteriorated.

The Trump administration previously positioned Saudi Arabia, UAE, Qatar, and Oman as core partners in AI and actively sought their financial support. However, the war has weakened the economic resilience of these oil countries and threatened their ability to continue investing in U.S. AI enterprises.

High Debt, Accelerating Financial Risks

Financial risks in the AI industry stem not only from external shocks, but its internal business logic also raises concerns.

Mega data center operators such as Microsoft, Google, Meta, and Amazon invested a total of nearly $700 billion in AI in a single year. To raise funds, data center providers are leveraging heavily, including structural financing arrangements with private equity institutions like Blackstone, BlackRock, and Blue Owl Capital. In 2025, mega operators will issue debts totaling $121 billion, four times the average of previous years, and this is expected to grow significantly.

Brad Lipton, former senior adviser to the U.S. Consumer Financial Protection Bureau and current director of corporate power and financial regulation at the Roosevelt Institute, noted: “The current situation is reminiscent of certain precursors of the 2008 financial crisis. All kinds of market players are highly correlated—banks lend to private credit institutions, which then funnel funds into other fields, thus amplifying systemic risks.”

The business model of the AI industry also faces endogenous deflationary pressures. The largest cost item for data centers, advanced AI chips, depreciates rapidly due to accelerated iteration, making the underlying asset value of data centers as debt not stable.

Meanwhile, AI services are billed by token, and unit costs continue to decline as model capability improves. Kedrosky describes this as the “death spiral to zero:” Token prices plummet, simultaneously weakening the overall value that data centers can create.

Clear Risk Transmission, Crisis Is Not Overhyped

Private equity firms are facing dual pressures: On one hand, the valuation of their acquired software companies is under pressure due to the AI shock; on the other hand, recent bets on data center investments are also in trouble.

Institutions like Blackstone and Blue Owl have invested heavily in data center construction with tech company rental income as a prerequisite for debt repayment, but as the cash flow of mega operators becomes tighter, the commercial viability of this model is being questioned. The funds for these private equity firms mainly come from pensions, endowments, and insurance institutions, thus extending the risk transmission chain to the entire financial system.

According to The Atlantic’s summary, even small problems in a few segments could simultaneously trigger a systemic crisis. For example:

  • Cost-side shock: Persistently high oil and gas prices drive up chip manufacturing and data center operational costs.
  • Industry funding chain fracture: Mega data center operators, already cash-strapped, are unable to pay rent; private credit institutions, also struggling, suffer heavy losses as AI-related bonds become “dead debt.”
  • Financial market transmission: Tech company valuations fall, dragging down public market performance; private equity institutions are forced to sell assets, putting tremendous pressure on institutional investors and banks.
  • Real economy imbalance: Years of excessive concentration of funds into data centers have led to a lack of investment in other economic sectors, leaving the overall economy weak.
  • Stagflation risk emerges: Ultimately, unemployment rises, and interest rates climb in tandem.

“The bursting of the bubble is inevitable; it’s the law of the economic system,” Lipton said. “What shouldn’t happen is the bursting of the bubble destroying the entire financial system. But what’s worrying is that the impact of AI investment is not limited to that; it may spread to the whole economy.”

Of course, a full-blown crisis is not the only possible outcome. Data center spending might cool down smoothly enough to avoid a hard landing; both Anthropic and OpenAI have achieved doubling revenue growth annually, and supporters believe generative AI products will ultimately turn profitable. But along the current trajectory, this goal will still take several years, and the risk of growth slowing or stagnating is equally real.

Risk Disclosure and DisclaimerThe market is risky, and investment needs to be cautious. This article does not constitute individual investment advice, nor does it take into account individual users’ specific investment goals, financial situation, or needs. Users should consider whether any opinions, viewpoints, or conclusions in this article fit their particular circumstances. Investing based on this is at your own risk.