Citi projects Alphabet, Meta and Amazon capex of $801 billion combined in 2027, highlighting AI infrastructure arms race

Capital spending by major tech firms is set to accelerate again, according to a Citi projection cited in a report published by Yahoo Finance. The bank estimates that Alphabet, Meta Platforms and Amazon together would spend $801 billion in 2027, a level that would dwarf what most companies have historically invested in aggregate in a single year.

The Citi view, as summarized in the report, ties the surge to the infrastructure demands of artificial intelligence workloads. AI systems require large amounts of compute capacity and, increasingly, the supporting power and cooling at data centers, which can turn multi-year technology roadmaps into large, forward-weighted spending plans.

For Alphabet, the market implication is straightforward but significant: any sustained increase in spending on AI-related hardware and capacity can affect near-term free cash flow even if it supports longer-term revenue growth. Alphabet, like other AI-heavy players, must balance the timing of infrastructure buildouts with the pace at which AI services can be deployed and monetized.

Meta Platforms and Amazon, included in Citi’s combined figure, face parallel constraints. Meta is building and running AI systems for products that rely on recommendations, content ranking and ad targeting, while Amazon’s AWS business supplies cloud compute demand. In both cases, increased capacity is partly about meeting customer demand and partly about improving the cost efficiency of running models at scale.

While the projected total is eye-catching, the report does not provide a company-by-company breakdown in the information available here, nor does it spell out how Citi allocates spending between data centers, networking, semiconductors, and other components. That means investors will need further clarity, either from the companies’ own capital expenditure guidance or from more detailed sell-side assumptions, before translating the combined number into expectations for each firm’s cash-generation profile.

The broader sector context is that AI spending is shifting from experimentation toward industrial scale. Over the past year, investors have increasingly focused on how quickly major platforms can convert rising infrastructure costs into usage, enterprise contracts, and advertising or cloud consumption. In that framework, a large 2027 capex projection suggests the AI buildout is expected to remain a central feature of technology company budgets for multiple years, not just a short-term cycle.

One caveat is what is not disclosed in the available reporting. Citi’s forecast methodology, assumptions about supply chain constraints, and the extent to which spending could be offset by efficiencies are not detailed in the excerpt available here. Likewise, there is no indication in the provided material of whether the $801 billion figure reflects committed spending, management targets, or scenario-based estimates.

For what to watch next, the key indicates will be each company’s own capital expenditure commentary, including any updates tied to data-center capacity, custom silicon and AI hardware roadmaps, and cloud utilization. If Alphabet’s results show that higher spending is translating into stronger cloud and AI product adoption, the market may become more tolerant of the cash impact; if not, the spending cycle could become a bigger earnings risk.