Iran War Impact on $2 Trillion AI Investment, Infrastructure and Hardware Hit First

The Middle East conflict is reshaping the AI investment landscape, tearing apart what once seemed like an ironclad tech feast into two completely different halves.

According to Bloomberg columnist Pam Olson, funds from the Middle East once pledged as much as $2 trillion, serving as a key foundation for the US AI race. However, as the US becomes involved in this conflict, the investment outlook is now clouded. Meanwhile, rising energy costs are significantly increasing data center operating expenses, and the chip supply chain is under pressure due to instability in the Middle East.

The ripple effects of this conflict are more likely to divide the AI market into two rather than destroy it entirely. She notes that hyperscalers like Alphabet, Amazon, and Microsoft are the most affected, while AI application companies like OpenAI and Anthropic are less impacted.

AI Software Layer: Enterprise Contracts Build Moats

Pam Olson points out that the AI market actually consists of two very different parts: one is the costly infrastructure business, and the other is the relatively low-cost software business.

In the software layer, Anthropic’s annualized revenue has more than doubled in the past three months to $19 billion; OpenAI’s annualized revenue is about $25 billion. Enterprise clients, consumers, and government agencies in industries like finance and life sciences continue to pay for subscriptions and access rights. Unlike the previous hype cycles around the metaverse and cryptocurrencies, this growth is considered more sustainable.

AI software companies also benefit from strong customer stickiness. Clients are less likely to cancel contracts due to geopolitical uncertainties and are more inclined to maintain subscriptions to improve organizational efficiency and hedge against potential economic fluctuations.

Additionally, AI software companies, while reliant on data center operations, are not directly bearing the brunt of rising energy costs. OpenAI and Anthropic mainly depend on “inference”—running existing models to respond to user queries—which consumes far less energy than training new cutting-edge models. Training the latter requires thousands of GPUs running continuously for weeks or even months, making it extremely energy-intensive and delaying progress.

Hyperscalers: Energy and Capital Under Double Pressure

In contrast, hyperscalers like Amazon, Alphabet, Microsoft, Meta, and Oracle are in a more vulnerable position. Pam Olson notes that these companies have collectively invested over $1.15 trillion in infrastructure, a massive expenditure highly dependent on cheap and stable energy supplies, especially natural gas.

According to the International Energy Agency, natural gas accounts for about 40% of electricity used by US data centers, making it the primary single energy source. Rising energy prices due to the Middle East conflict directly increase operational costs for these companies.

However, Alphabet and Amazon still have ongoing cloud subscription revenues as a financial buffer.

Chip Supply Chain: TSMC and Nvidia Face Highest Risks

The chip supply chain is also highly exposed to the Middle East situation. TSMC manufactures nearly all of Nvidia’s high-end chips, and about one-third of its fuel comes from the Middle East, with most of its helium supply sourced from Qatar. Helium is critical in semiconductor manufacturing for cooling and protecting silicon wafers. According to Olson, after last week’s drone attack on Qatar’s Ras Laffan industrial city by Iran, helium production has been constrained, and full chip capacity recovery could take months.

Within this landscape, Nvidia may face the highest risk. This $4 trillion market cap company’s revenue mainly comes from selling chips to hyperscalers. Any slowdown in large data center construction could directly impact its orders.

Unlike Alphabet and Amazon, Nvidia has no recurring revenue streams as a buffer and relies solely on chip sales to sustain its business. It faces a double blow: on one side, manufacturing is hindered; on the other, large orders previously secured in the Middle East are now uncertain. In November last year, the US government approved Nvidia’s sale of 70,000 advanced chips to the UAE and Saudi Arabia, but this deal now appears more uncertain. Nvidia declined to comment.

Olson emphasizes that energy resources and capital from the Gulf region have long been a driving force behind the boom in AI infrastructure. No matter how strong the revenue growth at the application layer, as long as the war continues, the prospects for underlying infrastructure will become increasingly fragile.

Risk Warning and Disclaimer

Market risks exist; invest cautiously. This article does not constitute personal investment advice and does not consider individual users’ specific investment goals, financial situations, or needs. Users should evaluate whether any opinions, viewpoints, or conclusions herein are suitable for their particular circumstances. Invest at your own risk.