The $720 Billion Capex Trap: 2 Artificial Intelligence (AI) Hyperscalers Spending on Growth While the Rest Spend on Maintenance

In 2026, the top five U.S.-based hyperscalers -- Microsoft (NASDAQ: MSFT), Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG), Meta Platforms (NASDAQ: META), Oracle (NYSE: ORCL), and Amazon (NASDAQ: AMZN) -- have projected that they will collectively spend a staggering $720 billion in capital expenditures. As aggressive as this figure appears, this phase of accelerating artificial intelligence (AI) infrastructure growth marks a moment during which the technology shifts from aspirational experiments to being a backbone of the global economy.

Industries are rapidly demanding intelligent systems that can learn, reason, and act at machine scale. The hyperscalers acknowledge that whoever controls the underlying infrastructure will likely capture the lion's share of AI-driven value in the coming decade.

While the race is fast-paced, not all participants carry equal conviction or clarity. Based on the catalysts propelling AI infrastructure build-outs, and the concrete use cases around these growing budgets, I see Microsoft and Alphabet as uniquely equipped to justify their commitments while the rest of big tech risks overextension.

Image source: Getty Images.

Why are AI hyperscalers accelerating infrastructure budgets?

AI capex budgets are a function of a simple reality: Appetite for AI computing power is growing at an incredible rate. Creating a generative AI model requires training sessions measured in millions of GPU hours, while inference demands scale exponentially as adoption of those models deepens across consumer and enterprise environments.

Companies are no longer considering whether or not to adopt AI, but rather how quickly they can embed new workflows into their core operations. This creates a feedback loop in which the highest-capable models unlock new use cases -- requiring developers to access critical infrastructure.

Hyperscalers that hesitate to invest heavily in new data centers risk becoming more of a utility in a landscape where differentiation will hinge on which providers can deliver the most advanced services at the lowest marginal cost.

When any of the players announces a breakthrough model or a new commitment of GPU clusters, the others are essentially forced to match or surpass that rival to avoid customer migration.

Breaking down the capex

The roughly $720 billion of AI infrastructure spend is not being allocated toward abstract research and development work or to marketing campaigns. It will largely be dumped into steel, silicon, and electrons.

The largest share will fund the construction of factories purpose-built for AI workloads -- data centers that eclipse traditional cloud campuses in power density and cooling sophistication. Inside these facilities are rows of liquid-cooled server racks housing hundreds of thousands of GPU clusters, interconnected by ultra-low latency fabrics.

Power infrastructure will consume another sizable portion of the expense stack. AI training clusters draw loads of electricity, forcing hyperscalers to commit to long-term agreements for renewable and nuclear capacity.

In addition, big tech is increasingly spending on designing proprietary silicon. These custom application-specific integrated chips (ASICs) allow companies to migrate beyond the GPU supply bottleneck and tailor chips to the workloads they will be handling.

Why Microsoft and Alphabet are better positioned than their peers

In my view, Microsoft and Alphabet stand apart from the competition because their AI infrastructure spending is tightly aligned with defensible, high-margin application layers that already touch hundreds of millions of users and enterprises every day.

Against this backdrop, their respective investments represent classic growth capex -- capital deployed aggressively to capture market share, accelerate revenue trajectories, and compound competitive moats. By contrast, the spending by their rival platforms carries a heavier flavor of maintenance capex. It is largely about sustaining existing footprints and defending market share rather than igniting near-term growth engines -- with payoffs that feel more distant and uncertain.

Microsoft's cloud platform, Azure, benefits from an unparalleled distribution channel: Microsoft Office, the world's most ubiquitous productivity suite. When Copilot adds new features within Word, Excel, and Teams, every enterprise license becomes a vector for AI consumption. This integration turns capex into revenue visibility, as customers are already paying for the applications and willingly pay a premium for AI layered on top.

Alphabet enjoys a similar advantage. Its Google Search, YouTube, and Android ecosystems generate one of the richest proprietary data streams in the world. Meanwhile, DeepMind's research pedigree and Google's custom Tensor Processing Units (TPUs) deliver efficiency edges that competitors cannot easily replicate at scale.

For now, Meta's AI ambitions remain focused on advertising optimization and wearable hardware experiments. Social platforms inherently face user fatigue issues and regulatory headwinds. Pouring billions of dollars into infrastructure to power recommendation tweaks or virtual reality and gaming features risks becoming more of a defensive upkeep play rather than an offensive expansion strategy.

Oracle operates from an even narrower base. Its cloud infrastructure presence, while growing, lacks the breadth of incumbents like Azure or Amazon Web Services (AWS). Furthermore, its database-centric history risks leaving portions of new AI capacity underutilized if clients decide to migrate workloads toward more general-purpose platforms.

Amazon's cloud investments compete internally with its core e-commerce business. Moreover, the company's customer relationships, while vast, lack the same level of application-layer lock-in that Microsoft and Alphabet enjoy.

Lacking a comparable proprietary model ecosystem like Google Gemini or a daily productivity hook like Microsoft Office, Amazon risks spending on new capacity where the returns on those investments are diluted by slower integrations against less certain demand -- more maintenance of an established foundation than bold growth into the next architecture.

In the end, I think Microsoft's and Alphabet's spending is justified because it reinforces flywheels that are already spinning at full speed across data, customers, distribution networks, and innovation. The other hyperscalers may ultimately find themselves spending on infrastructure simply to ride the rails of the AI economy as opposed to building it.

Adam Spatacco has positions in Alphabet, Amazon, Meta Platforms, and Microsoft. The Motley Fool has positions in and recommends Alphabet, Amazon, Meta Platforms, Microsoft, and Oracle. The Motley Fool has a disclosure policy.