AI capex to exceed half a trillion in 2026: UBS

UBS significantly increased its global artificial intelligence capital expenditure forecasts, now projecting spending to reach $423 billion in 2025 and $571 billion in 2026, culminating in $1.3 trillion by 2030 with a 25% CAGR. This upward revision is attributed to robust fundamentals, accelerating demand for computing power, and recent megadeals, underscoring that compute demand is outpacing expectations and monetization is accelerating. The bank suggests AI-related stocks will drive equity markets, recommending investors add diversified exposure to the theme given major U.S. tech companies' strong financial health despite rising capex intensity.

Analysis

UBS has significantly upgraded its global Artificial Intelligence (AI) capital expenditure forecasts, now projecting spending to reach $423 billion in 2025 and $571 billion in 2026, up from previous estimates of $375 billion and $500 billion, respectively. The bank anticipates overall AI capex to hit $1.3 trillion by 2030, implying a robust 25% compound annual growth rate over the next five years. This upward revision is primarily driven by robust fundamentals and an accelerating demand for computing power, with compute demand notably outpacing prior expectations.

The increased momentum is underscored by recent megadeals, such as the $38 billion, seven-year agreement between OpenAI and Amazon, and Microsoft's $9.7 billion commitment to acquire computing capacity from Australian data center operator IREN. These transactions highlight the escalating need for advanced computing infrastructure to support increasingly complex AI applications. Furthermore, data indicates a dramatic surge in AI token consumption, exemplified by Google's Gemini reporting a 130-fold increase over the past 18 months, alongside Meta's expanding compute requirements.

Despite the rising capital expenditures, major U.S. tech companies are noted for their continued financial strength, with capex intensity nearly doubling to 20.8% over the last five years and projected to reach 27% by 2030. UBS concludes that AI-related stocks are poised to be a primary driver of equity markets. The bank explicitly advises under-allocated investors to gain exposure to the AI theme, advocating for a diversified investment approach.

Key Decisions for Investors

• Investors under-allocated to the AI theme should consider increasing exposure through a diversified approach, given UBS's significantly upgraded capex forecasts and the expectation for AI-related stocks to drive equity markets.
• Focus on companies providing core AI infrastructure and computing capacity, as evidenced by megadeals like Microsoft's investment in IREN and the accelerating demand for compute power.
• Monitor the financial health and capex intensity of major tech companies, as projected increases to 27% by 2030 could impact future free cash flow generation and profitability.