Alphabet commits $811 billion（約130兆円） more to AI after $45 billion（約7.2兆円） Q2 spend

Alphabet has committed to one of the largest capital spending programs in corporate history, driven almost entirely by its race to build out artificial intelligence infrastructure. In its second-quarter earnings report, the company disclosed that it spent $45 billion（約7.2兆円） on capital expenditures in the quarter alone—double what it spent in the same period a year earlier. More dramatically, a footnote in its 10-Q filing with the Securities and Exchange Commission revealed that Alphabet has signed contractual commitments totaling $811 billion（約130兆円） as of the end of June, predominantly for AI-related infrastructure. This represents a staggering increase from just $332 billion（約53兆円） in commitments at the end of the first quarter.

These long-term purchase agreements are spread across three main categories: securing chip supply, funding data center construction, and locking in energy services to power those facilities. The company structured these as take-or-pay contracts—arrangements that guarantee either the delivery of goods or services or payment of a penalty—to secure guaranteed supply or favorable rates years into the future. Alphabet disclosed that it expects to generally fulfill all non-energy agreements by 2030, while energy service contracts range from two to 26 years, with some obligations extending through 2054. The company explicitly stated in its 10-Q that it is facing a "severe shortage of compute capacity" as it takes on massive multi-year customer deals, forcing it to pursue capacity increases through third-party providers as a temporary bridge until it can build out sufficient internal infrastructure.

The scale of these commitments has already altered Alphabet's financial profile. For the first time since its initial public offering in 2004, the company reported negative free cash flow in the second quarter, a direct result of the $45 billion（約7.2兆円） capital expenditure. To manage this, management raised Alphabet's full-year 2026 capital expenditure budget to between $195 billion（約31兆円） and $205 billion（約33兆円） and signaled that capex will "increase significantly in 2027." Despite the near-term cash flow pressure, company leaders framed the spending as a rational investment in a genuine market opportunity. The company is also pursuing an additional revenue stream by ramping up direct sales of its custom Tensor Processing Unit (TPU) systems—proprietary AI chips designed in-house. Supporting this effort, Alphabet's inventory jumped from $2.4 billion（約3800億円） to $10 billion（約1.6兆円） in the most recent quarter, reflecting both the scale of TPU production and the company's commitment to supply third-party customers with its silicon as it fulfills its long-term supply agreements.

Alphabet's massive AI spending commitment reflects a fundamental shift in how the company is managing its capital allocation. The disclosure of $811 billion（約130兆円） in contractual obligations—a jump of nearly $500 billion（約80兆円） in just one quarter—signals not only aggressive near-term investment but also a multi-year bet on sustained, growing demand for AI compute. By entering into long-term take-or-pay contracts for chips, data centers, and energy through at least 2030 (and energy services to 2054), Alphabet is locking in supply and pricing certainty even as it acknowledges severe capacity shortages today.

The shift into negative free cash flow, a first since the company's 2004 IPO, is notable but not necessarily alarming to management. Alphabet has disclosed that its core operations remain cash cows and its cloud business is generating strong returns on invested capital. The company is also hedging near-term capacity constraints by increasing third-party provider usage as a bridge until it can build out more internal capacity—a move that will create a short-term margin headwind but is framed as a long-term strategic necessity. Additionally, the jump in inventory from $2.4 billion（約3800億円） to $10 billion（約1.6兆円） and the ramp in direct Tensor Processing Unit sales suggest Alphabet sees opportunity not just in consuming AI compute internally but in selling it to external customers, which could help justify the enormous supply-chain commitments it has made.

No comments yet. Be the first to share your thoughts!