Alphabet is due to report second-quarter earnings after the U.S. market closes on July 22, or early July 23 in Beijing. The main question for investors is not quarterly profit. It is whether management will raise its AI capital expenditure guidance again.
2026 capex guidance may move higher
Robert Castellano, president of The Information Network and a researcher with more than 40 years of experience in the technology industry, wrote that multiple supply-chain signals point to Alphabet’s earlier 2026 capex outlook of $180 billion to $190 billion as potentially too conservative.
He said spending on AI infrastructure could keep climbing over the next several years as AI data center construction accelerates and cloud demand stays strong. Castellano expects Alphabet could raise its 2026 capital spending guidance to $190 billion to $200 billion in this earnings release, while leaving unchanged its commentary that 2027 spending will be close to $300 billion.
Deutsche Bank has also lifted its projections for Alphabet’s future capex. According to a previous Wallstreetcn report cited in the article, the bank now expects Alphabet’s 2027 capex to reach about $325 billion, up sharply from an earlier forecast of $250 billion. For 2028, it sees spending increasing further to $365 billion to $370 billion.
If that path materializes, it would indicate that the AI infrastructure investment cycle is still expanding. It would also force the market to reassess earlier concerns that AI capital spending was close to peaking and that growth momentum was starting to slow.
Five tech giants could spend more than $770 billion in 2026
The report argues that AI infrastructure investment is entering a much larger expansion phase, and that Alphabet’s capex trajectory now matters well beyond a single quarterly earnings release.
As hyperscale cloud companies keep putting more money into data centers, computing capacity and energy infrastructure, capital spending is shifting from gradual growth to a long-duration contest measured in hundreds of billions of dollars.
Market data cited in the article shows combined capex at Alphabet, Amazon, Microsoft, Meta and Oracle rose from $271 billion in the trailing 12 months through April 2025 to $482 billion in the trailing 12 months through April 2026. Based on company guidance and market expectations, full-year 2026 capex for the five companies is projected at $745 billion to $775 billion, up about 55% to 61% from the prior period.
That number reflects not only a bet by large technology companies on AI commercialization, but also their growing role as the core driver of demand across the semiconductor supply chain.
The spending extends beyond chips
The article notes that more than $770 billion in capex would not flow directly to chip suppliers alone. The total also covers servers, networking equipment, data center construction, land, power facilities and broader energy supply systems.
Still, each new AI data center feeds demand across several parts of the chain at the same time. That includes high-performance computing chips, HBM memory, advanced packaging, wafer manufacturing, test equipment and high-speed interconnect networks. In that sense, AI infrastructure spending is creating a broad pull effect across the full supply chain.
The market focus is that demand in the current AI supply chain is not being set by one company. It is being shaped by a small group of hyperscale cloud operators. Their purchasing pace for GPUs, custom ASICs, HBM, optical networks and data center power systems has become an important gauge for the health of the broader semiconductor cycle.
That is why any further increase in Alphabet’s capex matters not only for Gemini and Google Cloud. It is also a test of how strong the next phase of AI infrastructure demand may be.
Meta and Alphabet show two different capex paths
Meta’s recent consideration of selling part of its AI infrastructure capacity to external customers has sparked debate over whether AI investment may be running ahead of demand. The article says the move looks more like an attempt to improve asset utilization than a cut in investment.
Based on revised estimates from Wells Fargo, Meta’s decision to externalize part of its AI infrastructure capacity is mainly tied to a timing gap in the release of internal inference demand. Renting that capacity out would allow the company to generate returns earlier on assets it has already built. In other words, Meta is commercializing deployed compute resources rather than being forced to digest excess capacity because of weak demand.
Alphabet, by comparison, is described as having stronger internal absorption capacity. Google Cloud has a backlog of contracts totaling $462 billion, and enterprise demand for AI compute and cloud services already provides meaningful order support. That suggests new capital spending at Alphabet can map more directly to future revenue growth, giving its capex expansion a higher degree of commercial visibility.
A key readout for the AI cycle
For the market, the critical part of Alphabet’s July 22 earnings release is not just revenue and profit. It is whether management raises its expectations for AI capital spending.
If Alphabet lifts its 2026 capex guidance, that would strengthen the market view that the AI infrastructure investment cycle still has room to run, while offering fresh fundamental support for chips, memory, advanced packaging and the data center supply chain.
If spending comes in below expectations, the market may revisit assumptions about the payoff period for AI investment and put short-term pressure on valuations across related technology stocks.
With Microsoft, Meta and Amazon set to report after Alphabet, Wall Street will be watching whether the AI infrastructure race is still accelerating and whether hundreds of billions of dollars in investment can continue to convert into revenue growth.