The $1 trillion artificial intelligence (AI) market promises a revolution in business and technology, but its rapid growth could face significant threats, including supply chain overload and energy challenges. IT-World examines the key risks and prospects.
Last week, Alphabet CEO Sundar Pichai stated, “The risk of under-investing far outweighs the risk of over-investing.” And he wasn’t joking: Alphabet alone plans to increase capital expenditures by 50% this year, up to $48 billion, with most of that going toward AI equipment.
Analysts at New Street Research estimated that Alphabet, Amazon, Meta, and Microsoft together will spend $104 billion on building AI data centers this year. Including the expenses of smaller tech companies and other industries, total investments in AI data centers from 2024 to 2027 could reach $1.4 trillion. However, such massive spending is causing concern among shareholders. The day after Alphabet announced its results, the Nasdaq index dropped by 4%—the largest one-day decline since October 2022. Microsoft also announced plans to increase spending on AI infrastructure, which already led to a decrease in the corporation’s stock value. Analysts are now waiting for Amazon’s results to gauge the state of its AI business.
Meanwhile, tech giants continue to ramp up investments, benefiting numerous suppliers. Nvidia, a chip manufacturer for AI, briefly became the world’s most valuable company. But the AI supply chain is much broader, involving hundreds of companies—from Taiwanese server manufacturers to Swiss engineering firms and American utility companies. Half of all AI investments are directed at chip manufacturers, with Nvidia being the main beneficiary. The rest is spent on equipment that supports chip operations—from networking gear to cooling systems. In 2023, the average stock price of companies in this sector increased by 103%, compared to a 42% increase in the S&P 500 index. Expected sales for 2025 have also risen by an average of 14%, compared to 1% for other S&P 500 companies.
The greatest successes have been recorded by chip and server manufacturers. Nvidia expects to sell $105 billion worth of chips and related equipment this year, double the amount from the previous fiscal year. AMD, Nvidia’s closest competitor, is likely to sell around $12 billion worth of chips for data centers, up from $7 billion previously. Broadcom reported a 280% year-over-year increase in quarterly AI revenue, reaching $3.1 billion.
Server manufacturers are also seeing revenue growth. Dell and Hewlett Packard Enterprise reported that sales of AI servers doubled in the last quarter. Taiwanese manufacturer Foxconn, which assembles iPhones for Apple, also reported a threefold increase in AI sales over the past year.
However, alongside the growth in investments, there are threats to the supply chain. One problem is the high dependence on Nvidia. The future sales of many companies hinge on meeting the demands of this leading chip manufacturer. Another threat is bottlenecks in supplies, particularly in the area of energy. Bernstein’s analysis shows that if AI usage by 2030 reaches levels comparable to current Google Search usage, it could increase energy consumption in the U.S. alone by up to 7% annually.
Companies are already taking steps to address these issues by offering autonomous power sources. In March, Talen Energy sold Amazon a data center connected to a nuclear power plant for $650 million. CoreWeave, a small cloud AI provider, struck a deal with Bloom Energy to produce on-site energy. However, the most serious threat to the AI supply chain could be a decrease in demand. In June, Goldman Sachs and Sequoia published reports questioning the benefits of current generative AI tools. If AI profits remain modest, tech giants may cut capital expenditures, leaving the supply chain vulnerable.
AI continues to attract significant investment, promising revolutionary changes in the economy and business. However, the rapidly growing supply chain faces risks of overload, dependence on key suppliers, and energy challenges. The stakes for companies in the AI supply chain are rising, and questions about their resilience to potential market shifts remain open.