UBS Report: AI Capital Spending to Hit $1.2 Trillion Despite Slowdown Calls
A new UBS report analyzes global market trends, including artificial intelligence investment forecasts reaching $1.2 trillion by 2027, alongside rising U.S. Treasury yields crossing the 5% threshold.
A new report from UBS addresses the two main issues currently occupying global markets: concerns over a slowdown in the pace of artificial intelligence development and its potential impact on chip companies, alongside pressure in the U.S. Treasury market ahead of the Federal Reserve's interest rate decision.
UBS discussed the recent declines in AI-related stocks after several leading U.S. labs called for slowing the development pace of "Frontier AI" models and tightening safety standards. The Philadelphia Semiconductor Index (SOXX) fell 5.9% on Monday, and memory stocks in North Asia also saw significant declines, though market pressure eased on Tuesday.
AI Investment Trends and Forecasts
However, UBS emphasizes that the calls for slower development do not mean the AI capital expenditure cycle is ending. According to the review, about two-thirds of AI computing demand comes from running existing models—inference—rather than training new ones. Therefore, continued investment depends largely on organizational adoption of the technology and the ability to monetize its use.
The bank notes that a "slowdown" in development is not necessarily a halt to model training. Elon Musk recently confirmed that training for the Grok 4.8 model is ongoing, while OpenAI expects continued growth in memory demand. Accordingly, UBS maintains its forecast that AI capital expenditures will reach $1.2 trillion in 2027, compared to $900 billion this year—a 33% increase.
"Calls for stricter regulation may also be an attempt by major AI companies to influence future rules that could benefit established players while hindering startups or foreign competitors," the report suggests.
Market Valuations and U.S. Bond Yields
Recent declines have also shifted AI stock valuations. The SOXX index is now trading at a forward price-to-earnings multiple of 21, compared to 33 in June and an average of 24 since the launch of ChatGPT. The "Magnificent Six" are also trading at a multiple of 23, about 25% lower than their October valuation.
Meanwhile, U.S. bond yields are in focus. On Monday, the yield on the 10-year U.S. Treasury note crossed the 5% threshold, reaching 5.01% for the first time since 2023. UBS attributes the trend to inflation concerns and expectations of a prolonged high interest rate environment. Investors are now awaiting the Fed's decision and economic outlook.
