Google CEO Sundar Pichai made a statement that broke the typical corporate rhythm of an earnings presentation during the over hour-long Q1 2026 earnings conference in late April: “We are compute constrained in the near term.” If we could have satisfied the demand, our cloud revenue would have increased.
That is an impressive statement from the head of one of the world’s most capital-intensive businesses, which just guided to between $180 and $190 billion in capital expenditures for 2026—nearly twice as much as it spent in 2025—and then raised $84.75 billion in debt, equity, and private placement to finance the expansion. There is a demand. The infrastructure isn’t big enough to support it yet. Alphabet currently resides there.
Due to oversubscription, the stock offering, which priced on June 3, 2026, was increased from the original $80 billion target. It consists of a $40 billion at-the-market program, $30 billion in publicly underwritten tranches, and a $10 billion private placement by Berkshire Hathaway, which garnered the most attention. Historically, aggressive technology bets on unproven infrastructure cycles have not been linked to Warren Buffett’s company.
The fact that Berkshire, which has been discreetly increasing its ownership of Alphabet since Q3 2025, committed $10 billion as an anchor investor at $351.81 a share, carries a particular weight in the investment community. It is not evidence that the capital investment will yield returns that are proportionate. It is proof that those who have made the most accurate long-term capital allocation judgments think the wager is reasonable.
The operational data point that makes the bull argument clear is Google Cloud’s Q1 2026 figures. At a run rate that puts the company above $70 billion yearly, revenue hit $20 billion in the quarter, a 63 percent year-over-year gain. The signed commitments that have not yet been turned into revenue, or the contracted backlog, almost doubled over time to reach over $460 billion. Within a year, more than half of the backlog is anticipated to be converted.
In discussions with investors, CFO Anat Ashkenazi frequently brings up the backlog figure because it reflects proven and committed demand as opposed to projected demand. Pichai’s compute constraint is not a demand issue. The issue is one of supply. The $190 billion invested in 2026 and whatever is spent in 2027 represents an effort to close the gap between what consumers desire and what is available.
It is instructive to observe how the balance sheet changes as a result of this investment program. While cash and marketable securities stayed at $126.8 billion, long-term debt increased from $46.5 billion at the end of 2025 to $77.5 billion by the end of Q1 2026. The company’s free cash flow for the first quarter was $10.1 billion, which is a respectable figure on its own but a substantially different picture than what investors have long believed.
These data centers’ eighth-generation Tensor Processing Units, which were specially created for Gemini, are an effort to vertically integrate the supply chain in a way that lowers long-term costs and reliance on outside chip vendors.
The quality of the platform Alphabet is developing is confirmed by the cooperation with Apple for Siri AI workloads, a corporation that is well-known for discriminating against its infrastructure partners. However, the energy consequences are real, and corporate sustainability teams are quietly putting in extra effort to solve Alphabet’s admitted slippage against its 2030 net-zero targets.
Reading the Q1 figures and the documentation that followed the equity offering gives the impression that Alphabet’s identity as an investment has evolved more in the last six months than it has in the previous five. This business is no longer buying back shares and producing a lot of free cash flow.
This company is conducting an exceptionally large-scale build program, clearly informing investors that 2027 would be significantly more expensive, and wagering that the contracted backlog turning into revenue will eventually validate numbers that appear exceptional in comparison to any historical comparison. They might be correct. The $460 billion backlog indicates that the market agrees.
