UBS reports AI infrastructure stocks surpass big tech hyperscalers

The biggest winners of the AI spending bonanza aren’t the companies writing the checks. They’re the ones cashing them.

UBS Global Wealth Management research indicates that AI infrastructure stocks, the energy companies, data center equipment makers, and industrial suppliers powering the AI revolution, have outpaced the big tech hyperscalers that are actually deploying the capital.

The capital expenditure arms race

The numbers driving this shift are genuinely staggering. UBS has raised its global AI capital expenditure forecast to $423 billion for 2025, with projections climbing to $571 billion or more in 2026. Some UBS estimates push the 2026 figure as high as $820 billion.

The hyperscalers, primarily Microsoft, Amazon, and Alphabet, are leading this charge. UBS projects their combined capital expenditures could exceed $600 billion to $725 billion in 2026 alone. More than 85% of historical AI capex is tied to the largest hyperscalers, making them the single most important demand signal in the entire technology supply chain.

These companies are building data centers at an unprecedented pace, buying chips by the warehouse-full, and scrambling for power capacity wherever they can find it. In many cases, this spending consumes a massive portion of their operating cash flows.

Follow the electricity

Every new data center needs enormous amounts of electricity, cooling systems, and grid connections. UBS predicts hundreds of billions in spending on power generation, storage, and grid infrastructure through 2030, all tied directly to data center growth.

Energy and power-related companies have already started reflecting this demand in their stock prices. Companies like Fluence Energy and Hut 8 have seen significant price movements linked to supply agreements with hyperscalers. When Microsoft or Amazon signs a long-term power purchase agreement, the counterparty’s stock tends to respond favorably.

UBS has flagged record order backlogs in power and infrastructure supply chains as evidence that this trend has legs. Recent UBS analysis suggests that profit-taking has already begun in the semiconductor and hardware AI space, with capital rotating into more defensive infrastructure segments.

What this means for crypto and digital asset investors

The AI infrastructure trade has direct implications for crypto markets, even if the connection isn’t immediately obvious.

First, the competition for power is intensifying. Bitcoin miners and AI data centers are now competing for the same scarce resource: cheap, reliable electricity. Companies like Hut 8 have already started straddling both worlds, operating mining rigs alongside AI compute facilities. As hyperscalers bid up the price of power purchase agreements, pure-play Bitcoin miners without diversified revenue streams could face margin pressure.

Second, the broader macro signal matters. When UBS warns of potential overheating in the AI sector, it’s worth paying attention. The wealth management giant is advising selectivity in investment strategies rather than blanket AI exposure.

Third, the sheer scale of capital being deployed into AI infrastructure creates opportunities for tokenized real-world asset plays. Data centers, power plants, and grid infrastructure are exactly the kinds of assets that DeFi protocols are increasingly looking to bring on-chain. A $423 billion capex market in 2025, potentially doubling by 2026, represents a massive addressable market for tokenization platforms.

For investors watching the AI-crypto convergence, the key metric to track is hyperscaler capex guidance during earnings calls. When Microsoft, Amazon, or Alphabet revise their spending plans upward, it creates a cascading demand signal through energy, cooling, semiconductor, and networking supply chains.

UBS’s emphasis on diversifying into AI-adjacent infrastructure rather than concentrating in mega-cap tech suggests that the smartest institutional money is already making this pivot.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.