Google Cloud starts shipping TPU systems, becomes chipmaker

Google Cloud started recognizing sales of TPU system sales in the second quarter for the first time, which helps explain the 82% surge in revenue to $24.8 billion.

With those early TPU system sales, Alphabet is now an emerging semiconductor vendor. Scaling TPUs is another reason Alphabet raised its 2026 capital expenditure guidance to $195 billion to $205 billion.

TPU system sales are now reflected in Google Cloud's $514 billion in cloud backlog. Most of that backlog is for cloud agreements, but TPU system sales are included. Alphabet noted that it is building inventory to sell TPU systems. Google launched its eighth generation TPUs with TPU 8t and TPU 8i, one for training and one for inference.

TPU revenue will mostly be recognized in 2027. CFO Anat Ashkenazi said on Alphabet's second quarter earnings call:

"We started delivering TPU systems to customer data centers in the second quarter. We continue to expect to recognize a relatively small portion of the revenues from our existing TPU system sales agreements this year, ramping as we exit 2026.

We anticipate the vast majority of the revenues from these agreements will be realized in 2027. And given the supply constrained environment, we plan to expand the use of third-party capacity in Q3 as a bridging strategy while we build up more internal capacity."

Ashkenazi noted that Google Cloud is benefiting from its integrated stack approach that includes its own TPUs and hardware. From a business perspective, she said "the way to think about it is as an expansion of our total addressable market."

• Google Cloud presses full-stack AI edge with new TPUs, Agentic Data Cloud, Gemini Enterprise Agent Platform
• Anthropic inks deal with Google, Broadcom for TPU capacity
• Google Cloud Next 2026: A look at the big themes
• Google Cloud, AWS, Microsoft Azure: The AI vertical integration race

Until Alphabet fully ramps Google's TPUs for internal and external use, the company has to deploy an allocation strategy that prioritizes use cases. Alphabet CEO Sundar Pichai was asked about the allocation strategy.

Pichai laid out the following:

• The first priority is "making sure we are allocating what we need to compete at the frontier in terms of AGI development and so that is the foundation for everything we do."
• Given strong demand from cloud customers, Alphabet is using both TPUs and GPUs to serve models.
• The company is balancing TPU deployments headed to data centers. Pichai said it is making sure that it's allocating TPUs toward frontier model development.
• TPUs are also aimed at use cases for Alphabet's consumer and enterprise businesses leveraging first party models.

This TPU juggling act will continue for Alphabet's foreseeable future, but the big takeaway is that AI infrastructure system sales are going to become a real revenue line.