Alphabet falls as Google Cloud yields to AI

Alphabet stock declined 2% to $331.83 on Wednesday morning after CEO Sundar Pichai said Google Cloud is deliberately forgoing short-term revenue to free up computing capacity for frontier artificial intelligence research. The drop came even as the NASDAQ 100 gained 1%, leaving Alphabet underperforming other large technology names.

While Alphabet fell more sharply, Microsoft (MSFT) shares were down 0.3% to $530.24 and Amazon (AMZN) slipped 0.5% to $261.31. The move in Alphabet reflects investor focus on the financial impact of diverting cloud resources toward AI development at a time when overall demand for computing power is strong.

Pichai has repeatedly emphasized that Google is prioritizing frontier AI work over maximizing current cloud revenue. “Our cloud business is at any given moment in time forgoing short-term revenues because we are making compute trade-offs to support frontier AI development,” he said in a recent Big Technology Podcast interview.

He described ensuring Google DeepMind has sufficient compute to build cutting-edge models as “non-negotiable,” with other parts of the company planning around that constraint. Pichai characterized resource allocation to frontier AI as “ground zero” for Google, underscoring that research access to compute comes ahead of additional cloud sales.

The policy is already affecting prospective customers. Pichai said that taking on new clients requesting capacity for future years would be “incredibly difficult,” signaling that Google is capacity-constrained and rationing access even as commercial demand remains high.

The decision to sacrifice some cloud revenue comes against a backdrop of rapid growth in Google Cloud. Revenue for the unit surged 82% to $24.8 billion in the second quarter, and its contracted backlog reached nearly $514 billion, illustrating strong demand for its services despite the compute bottlenecks.

Alphabet is responding to the shortage with substantial capital spending. The company expects to invest between $195 billion and $205 billion in capital expenditures this year, with much of that directed toward expanding AI infrastructure such as data centers and specialized hardware.

Pichai linked today’s constraints to earlier decisions, saying current compute availability reflects choices made in 2024 and noting the long timelines required to bring new data centers online. He did not quantify how much cloud revenue is being deferred or lost, leaving the exact financial impact of the AI-first strategy unspecified.

The strain on resources is amplified by the launch of new AI models. Pichai said Gemini 4 Argon processed nearly 25 trillion tokens in its first 72 hours, roughly three times the initial usage of Gemini 3, and that Google is already capacity-constrained in serving the newer model.

With Alphabet scheduled to report earnings on October 28, investors will be watching for indications of how the company’s AI investment and capacity trade-offs are translating into financial performance, particularly within Google Cloud.