Alphabet raised its 2026 capital expenditure guidance, citing sustained shortages of Google Cloud computing capacity.1 CFO Anat Ashkenazi said the company remains in a "supply constraint environment," a condition she has now reiterated for multiple consecutive quarters.1
The constraint is not new, but its persistence is driving the spending decision. Ashkenazi's repeated acknowledgment signals capacity shortfalls have outlasted earlier expansion plans.1 Alphabet's response is to add more infrastructure rather than wait for existing buildouts to close the gap.
Alphabet is not spending alone. Four major AI hyperscalers plan roughly $650 billion in combined 2026 capital expenditure.2 That figure marks one of the largest coordinated infrastructure buildouts in the technology sector's history, spread across data centers, chips, and networking.
Demand data backs the capacity argument. CEO Sundar Pichai said demand for Alphabet's AI models is translating into strong token usage across both developers and enterprise customers.1 Rising token consumption is a direct measure of AI workload volume moving through Google's infrastructure, and it links customer demand to the capacity Ashkenazi says is already stretched.
The pattern raises a broader question for the other three hyperscalers competing for AI workloads. If Alphabet's capex increase follows disclosed capacity constraints rather than preceding them, the same sequence may play out across the industry as usage climbs.2 Tracking whether future capex revisions consistently follow reported utilization and backlog pressure, rather than anticipate it, will show whether this is an industry-wide pattern or specific to Alphabet.
The next two quarters of earnings calls will be the test. Capacity utilization commentary, backlog disclosures, and any further capex revisions from the four largest hyperscalers will show whether supply constraints are the consistent trigger behind the $650 billion spending envelope, or whether other factors are driving the buildout.2