Google is prioritising artificial intelligence research over some potential short-term cloud computing revenue as it competes to develop advanced AI models. CEO Sundar Pichai has acknowledged that the company is deliberately allocating computing resources to its AI research division, even when doing so limits the capacity available to paying cloud customers.

Pichai said Google Cloud was already giving up short-term revenue because of computing capacity trade-offs made to support frontier AI development. The comments highlight the growing pressure on technology companies to balance commercial cloud demand with the infrastructure required to build increasingly capable AI systems.

Google DeepMind Gets Priority Access to Computing Power

Pichai described providing sufficient computing capacity to Google DeepMind, the company's AI research division, as a non-negotiable priority.

Developing advanced AI models requires substantial computing power for training, testing and deployment. Google must therefore decide how to distribute its available chips, servers and data centre capacity between internal research projects and external customers purchasing cloud services.

The company believes that staying at the forefront of AI development could strengthen its broader technology business over the long term. Advanced models can support products ranging from Google Search and Gemini to enterprise AI services and developer tools.

However, prioritising internal research can limit the amount of computing capacity available to customers who want to run their own AI workloads on Google Cloud.

Google Cloud Demand Continues to Rise

Google Cloud's business is expanding rapidly as companies invest in AI infrastructure, cloud computing and enterprise applications.

According to the reported figures, Google Cloud revenue increased 82% to $24.8 billion in the second quarter, while its contracted backlog approached $514 billion. These figures indicate substantial demand for the company's cloud services.

Despite this growth, Pichai warned that accommodating new customers seeking computing capacity for 2027 would be extremely difficult.

The challenge reflects a broader industry shortage: demand for AI computing infrastructure is growing faster than providers can build and deploy new capacity.

Building AI Infrastructure Takes Years

Pichai explained that computing capacity available today reflects investment decisions made years earlier. Data centres require long planning and construction periods, while equipment procurement, electricity supply, land availability and specialised labour can all constrain expansion.

Google is increasing its infrastructure investment to meet expected demand. Alphabet's projected capital expenditure for 2026 stands at approximately $195 billion to $205 billion, much of it associated with expanding AI infrastructure.

Even with significant spending, additional capacity cannot be brought online immediately. This creates difficult decisions about which workloads receive priority while new infrastructure is being developed.

The Trade-Off Between AI Research and Cloud Revenue

Google's strategy reflects a fundamental question facing technology companies: should computing resources be sold to customers today or reserved for internal AI research that could create future competitive advantages?

Cloud services generate revenue directly, while AI research requires significant investment before its commercial returns become clear. By prioritising its own research, Google is betting that stronger AI capabilities will support future growth across its products and services.

The approach also carries risks. Restricting cloud capacity could delay customer projects or push businesses towards competing providers. Meanwhile, large infrastructure investments may take time to generate returns, particularly if AI demand or technology requirements change.

Pichai acknowledged the need to balance these priorities carefully, emphasising that the company intends to invest enough to remain competitive at the frontier of AI.

AI Compute Shortages Create Opportunities for Competitors

The shortage of computing capacity is creating opportunities for companies that can provide access to AI infrastructure. Cloud providers and specialised computing businesses are competing to supply the processing power required by AI developers.

Google's reported agreement to pay SpaceX approximately $920 million per month for computing capacity illustrates the scale of the demand and the importance of securing additional infrastructure.

For enterprise customers, the capacity crunch could mean longer lead times, higher infrastructure costs and a greater need to plan AI deployments in advance. Businesses may increasingly consider multiple providers or alternative computing arrangements to reduce their dependence on a single source of capacity.

What Google's Strategy Means for the AI Industry

Google's decision highlights how AI competition is increasingly shaped by access to infrastructure, not just model design and software development.

Companies with sufficient computing capacity can train and deploy advanced models more rapidly, while those facing shortages may struggle to meet customer demand or keep pace with competitors.

At the same time, the growing scale of infrastructure investment raises questions about long-term returns, energy consumption and the risk of building capacity that may eventually exceed demand.

For now, Pichai's comments suggest Google sees AI infrastructure as a strategic investment that warrants priority, even when that decision limits immediate cloud revenue.

Conclusion

Google's willingness to forgo some short-term cloud revenue underscores the importance it places on maintaining AI leadership. By prioritising computing resources for Google DeepMind, the company aims to develop advanced models that could strengthen its wider business over time.

The strategy comes with trade-offs, including constrained cloud capacity and substantial infrastructure costs. As competition intensifies, Google's ability to expand computing resources while satisfying external customers will be an important factor in determining whether its AI investments deliver sustainable long-term growth.