Google Major Earnings Night: Capex Guidance May Be Raised? Senior Analyst: Could Reach $200 Billion by 2026

Alphabet, Google's parent company, will announce its Q2 earnings after the U.S. market closes on July 22 (early morning of July 23, Beijing time). However, the market’s real focus is not on the single quarter profit performance, but whether management will further raise its AI capital expenditure (Capex) guidance.

Robert Castellano, president of The Information Network with over 40 years of technology industry research experience, pointed out that several industry chain signals indicate that Alphabet's previously given Capex forecast of $180–190 billion for 2026 may already be conservative. As AI data center construction continues to accelerate and cloud computing demand remains strong, Alphabet’s AI infrastructure investment over the next few years could continue trending upward.

Castellano expects that Alphabet may raise its 2026 Capex guidance to $190–200 billion in its earnings report, while maintaining the outlook that 2027 capital expenditure will approach $300 billion.

At the same time, Deutsche Bank has also substantially increased its forecast for Alphabet’s future capital expenditures. According to a previous “Wallstreet Insights” article, Deutsche Bank projects that Alphabet’s 2027 Capex will reach about $325 billion, a significant increase compared to the previous forecast of $250 billion; capital expenditure in 2028 could further rise to $365–370 billion.

If this trend materializes, it would mean that the AI infrastructure investment cycle is still expanding, and previous concerns about AI capital expenditures soon peaking or growth momentum slowing may need to be repriced.

AI Arms Race Heats Up: Top Five Tech Giants’ Capex Could Exceed $770 Billion in 2026

The wave of AI infrastructure investment is entering a stage of much larger scale expansion, and the trajectory of Alphabet’s capital expenditures now has broader market implications that go far beyond a single quarterly earnings report.

As hyperscale cloud providers continue to increase their investments in data centers, computing power, and energy infrastructure, capital expenditure is shifting from incremental growth to a long-term race measured in hundreds of billions of dollars.

According to market data, the combined Capex of Alphabet, Amazon, Microsoft, Meta, and Oracle has surged from $271 billion for the twelve months ended April 2025, to $482 billion for the twelve months ended April 2026. Based on current company guidance and market expectations, the five tech giants are projected to spend $745–775 billion in Capex throughout 2026, an increase of approximately 55–61% compared to the previous cycle.

This figure not only reflects big tech's bet on the commercialization prospects of AI, but also means they are becoming the core drivers of demand growth for the entire semiconductor supply chain.

Five Giants Dominate the AI Infrastructure Investment Cycle

It’s important to note that the more than $770 billion of capital expenditure will not all flow directly to chip suppliers. This figure includes spending on servers, network equipment, data center construction, land, power facilities and energy supply systems across multiple fields.

However, from a supply chain perspective, every additional AI data center will drive simultaneous expansion across multiple segments. From high-performance computing chips and HBM memory, to advanced packaging, wafer manufacturing, test equipment, and high-speed networking, AI infrastructure investment is creating a complete supply chain pull effect.

The core market focus is that current AI industry demand is not driven by any single company, but jointly determined by a handful of hyperscale cloud providers. Their pacing of purchases for GPUs, custom ASICs, HBM, optical networks, and data center power systems has become a key indicator of the overall health of the semiconductor cycle.

Therefore, whether Alphabet further raises its Capex is not only a judgment on the outlook for the Gemini model and Google Cloud business, but also a validation of the next phase of AI infrastructure demand intensity.

Meta Leases for Flexibility, Google Secures Orders—Two Paths for AI Capex

Recently, Meta's consideration of selling part of its AI infrastructure capacity to external customers sparked discussions about potential overinvestment in AI. But from an industry logic perspective, this move is more about improving asset utilization rather than cutting investment.

According to Wells Fargo’s adjusted forecasts, Meta’s drive to externalize part of its AI infrastructure capacity is mainly due to an internal timing gap in inference demand release, hoping to generate returns on assets in advance through external leasing. This means Meta is commercializing its already invested computing resources, not being forced to absorb excess capacity due to insufficient demand.

In contrast, Alphabet has a stronger internal absorption capability. Google Cloud currently has a backlog of contracts totaling $462 billion and enterprise clients’ robust demand for AI computing power and cloud services already provides strong order support. This means that Alphabet’s additional capital investment can more directly correspond to future revenue growth, making the business certainty of Capex expansion higher.

Alphabet Earnings Report Becomes a Key Window into the AI Cycle

For the market, the most critical aspect of Alphabet’s July 22 Q2 report is not just profit and revenue but whether management further raises its AI Capex expectations.

If Alphabet raises its 2026 Capex guidance, it will reinforce the market’s belief in the durability of the AI infrastructure investment cycle, providing new fundamentals support for chips, memory, advanced packaging, and the data center supply chain.

Conversely, if Capex comes in below expectations, the market may re-evaluate the AI investment return cycle and place short-term pressure on related tech stock valuations.

As Microsoft, Meta, and Amazon release their results subsequently, Wall Street will further observe whether this AI infrastructure arms race is still accelerating, and whether hundreds of billions in capital expenditure can continuously translate into revenue growth.