Alphabet’s Cloud capacity squeeze, and the capex plan to fix it, were visible long before its AI-era stock run

Alphabet’s latest surge in market attention has prompted investors to look back at how quickly the company was warning that AI-related demand was outpacing the supply of computing capacity. A Trefis analysis published Monday argued that the “good problem” behind Alphabet’s performance was not just adoption of its AI products, but a persistent capacity bottleneck that management had been flagging for months on earnings calls. (The piece frames the issue as customers wanting to buy, but Google not having enough infrastructure to serve everyone immediately.)

On Alphabet’s Q4 2024 earnings call, Chief Financial Officer Anat Ashkenazi tied Google Cloud growth to a supply-constrained environment. Ashkenazi said the company “ended the quarter at” $12 billion of Cloud revenue, up 30% year over year, and then described the operational reality behind that growth: “We exited the year with more demand than we had available capacity.” He added that the company was in a “tight supply-demand situation,” and that it was “working very hard to bring more capacity online.”

Management also quantified the spending commitment meant to relieve that constraint. In the same call, Alphabet said it expected to invest approximately $75 billion in capital expenditures (capex) in 2025, primarily for technical infrastructure such as servers, followed by data centers and networking. It also noted that roughly $16 billion to $18 billion of that amount was expected in the first quarter, while acknowledging that timing of deliveries and construction schedules could cause quarter-to-quarter variability.

Those capacity warnings were paired with indicators that Alphabet’s AI developer platform was scaling quickly. In prepared remarks during the Q4 2024 call, Alphabet said its Vertex AI developer platform saw a 5x increase in customers year over year, and that “Vertex usage increased 20x during 2024.” The company also said developers doubled to 4.4 million in about six months. Vertex AI is Google’s platform for building and deploying machine-learning models, and “usage” and “customers” are key measures of whether enterprises and developers are actually adopting the infrastructure-backed products.

By the following quarter, Alphabet was still describing Cloud demand as outstripping available capacity. In the Q1 2025 earnings call transcript, Ashkenazi reiterated the pattern, noting that the company had exited the quarter with more customer demand than capacity and that it wanted to ramp up to support customer needs and demand. In parallel, CEO Sundar Pichai highlighted accelerating demand indicates for Gemini, saying that “Active users in AI Studio and Gemini API have grown over 200% since the beginning of the year.” Gemini API refers to a set of application interfaces developers use to build services using Gemini models, typically charging per usage and tied directly to compute throughput.

The more recent investor messaging in June 2026 echoed the same theme, but on a larger scale. In a June 3 investor presentation tied to a proposed major equity raise, Pichai said Alphabet was experiencing “strong demand” for AI solutions “at levels that are meaningfully exceeding our available supply,” and that it intended to scale investments to secure foundational infrastructure. The presentation also projected that capex would expand sharply, with 2026 spend expected to be much larger than earlier years, and described funding plans combining operating cash flow, debt, and the equity offering.

Still, the public record around these calls leaves practical questions unanswered. Alphabet did not specify, in the earnings-call transcripts cited here, the exact magnitude of “turned away” business in dollar terms, or the precise schedule for when incremental capacity would fully eliminate the shortfall by product or customer segment. The company also did not fully break out how much of the constraint was attributable to specific bottlenecks such as data centers, servers, networking, or specialized AI chips, beyond describing technical infrastructure as the main target of capex.

What to watch next is how Alphabet matches its infrastructure buildout to revenue conversion. Investors are likely to focus on whether Google Cloud can sustain growth without continuing to describe supply limits as a central operating constraint, and whether Gemini-related products keep translating usage and developer adoption into recurring enterprise and consumer monetization. If management’s repeated pattern holds, the near-term risk is not that demand disappears, but that capacity takes longer to catch up than the market expects.