The AI Capex Trade: How Hedge Funds Can Access the Infrastructure Build-Out Beyond Mega-Cap Tech

The AI infrastructure build-out has moved from a thematic narrative to one of the largest sustained capital-deployment cycles in the history of corporate America. Aggregate capital expenditure by the four largest hyperscalers is approaching $700 billion in 2026, nearly double 2025 levels, with non-AI workloads accounting for a shrinking share. The investable surface for hedge funds is no longer concentrated in mega-cap tech equity. It now extends across power generation, grid infrastructure, semiconductor capital equipment, real estate, industrial gases, cooling and credit instruments tied to the build itself. The institutional opportunity is not in calling the next quarter's earnings beat. It is in expressing the multi-year capital cycle through positions diversified across the layers of the build.

"Mega-cap tech is the visible part of the AI capex trade. The investable surface is much larger and, in some segments, much less crowded. The hedge fund opportunity is in the layers of the build that allocators understand intuitively but most equity portfolios under-reflect: power, semis, real estate, gases, credit. A diversified expression of the capex cycle is the institutional version of what retail captures through a single Nasdaq position." David Lloyd, Chief Executive Officer of CV5 Capital

The Scale of the Build

The numbers are unprecedented in scale and unprecedented in speed of escalation. In 2024, the combined capex of the four biggest hyperscalers was just over $200 billion. Two years later, it is on track to approach $700 billion, with consensus 2026 estimates ranging from $660 billion to $725 billion depending on the analyst. Microsoft has lifted its capex outlook to $190 billion, with approximately $25 billion linked to rising component costs. Alphabet has revised its guidance to between $180 and $190 billion. Meta has increased its range to $125 to $145 billion. Amazon Web Services is expected to exceed $200 billion. Oracle is targeting approximately $50 billion. CreditSights estimates approximately 75 percent of aggregate hyperscaler capex in 2026 will fund AI-related infrastructure, representing approximately $450 billion in AI-specific spending.

The composition of the spend matters as much as the headline. Two thirds of Microsoft's recent quarterly capex went to short-lived assets, primarily GPUs and CPUs, against a depreciation profile materially shorter than traditional data centre infrastructure. The implication is a recurring capex stream rather than a one-time build, with replacement cycles compressed by the rapid advance of accelerator architecture.

The Six Investable Layers Beyond Mega-Cap Tech

The Cash Flow Risk That Defines Position Sizing

Hyperscaler free cash flow is being absorbed by the build at a pace that is changing the financial profile of the customers. Analyst estimates suggest Amazon's free cash flow could turn negative by approximately $17 to $28 billion in 2026, while Meta's free cash flow could decline by close to 90 percent. Microsoft's free cash flow is expected to slide approximately 28 percent before recovering in 2027. The implication for investors is twofold. First, the build is internally financed but at the cost of equity-holder returns in the near term. Second, any meaningful disruption to AI revenue trajectories produces an immediate question about whether the capex pace is sustainable.

The institutional response is to size positions in the supply-side trade against scenarios in which capex slows. A 20 percent reduction in 2027 capex would still leave hyperscaler spend at multiples of 2024 levels, but it would compress the tail of the build for some suppliers. Position sizing therefore reflects a bounded thesis rather than a permanent capex regime.

Operational Architecture for AI Capex Strategies

The institutional fund running an AI capex strategy combines equity, options and credit exposures across multiple sectors. The operational architecture must support that complexity:

• Cross-asset risk aggregation that captures the underlying single-factor exposure to AI capex, regardless of how the position is constructed.
• Sector concentration limits that prevent the fund from becoming a mono-thematic vehicle disguised as a diversified strategy.
• Independent valuation of private credit positions where applicable, with documented methodology consistent with fair-value accounting standards.
• Liquidity policy that recognises the differential liquidity of public equity, listed credit and private credit components, with redemption terms designed against the slowest leg of the portfolio.
• Disclosure that explicitly identifies the AI capex thematic exposure as a primary risk factor, with documented stress scenarios for capex slowdown.