Sundar Pichai Says Google Cloud Is Sacrificing Short Term Revenue to Support Advanced AI Development
Google CEO Sundar Pichai has revealed that the company is giving up potential short term revenue from its cloud computing business to support the development of advanced artificial intelligence models. The decision highlights Google's focus on maintaining its position in the rapidly evolving AI industry, even as demand for commercial cloud services continues to increase.
In an interview with Alex Kantrowitz on the Big Technology Podcast, published on October 20, 2026, Pichai explained that Google must carefully allocate its available computing resources between external customers and its internal AI research teams. He said providing sufficient computing capacity for frontier AI development remains a fundamental priority for the company.
The comments come at a time when technology companies are competing for access to powerful chips, servers, electricity and data centre infrastructure needed to train and operate increasingly sophisticated AI models.
Google Prioritises Frontier AI Research
Frontier AI refers to the development of highly capable artificial intelligence models that aim to push the limits of current technology. Building these systems requires significant computing power, specialised hardware and extensive research.
Pichai explained that Google considers access to computing resources essential for Google DeepMind, its AI research organisation. The company must therefore balance the commercial opportunity to sell computing capacity through Google Cloud against the need to use that capacity for internal research and model development.
“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,” Pichai said during the interview.
His remarks indicate that Google is prepared to forgo some immediate business opportunities when computing resources are needed to advance its AI capabilities. However, he did not disclose the exact amount of revenue the company expects to sacrifice through these allocation decisions.
The strategy reflects Google's broader ambition to develop advanced AI technology that can eventually support products and services across its business, including consumer applications, enterprise tools and cloud offerings.
Google Cloud Faces Strong Demand
Google Cloud provides computing infrastructure, data storage, software platforms and AI services to businesses and developers. Demand for these services has increased as organisations adopt AI applications and require more computing power to operate them.
According to figures cited in recent reporting on Alphabet's second quarter results, Google Cloud generated 24.8 billion dollars in revenue during the quarter, representing an increase of 82 percent compared with the same period a year earlier. Its contracted revenue backlog stood at approximately 514 billion dollars at the end of June.
These figures highlight the scale of demand for Google's cloud services. At the same time, allocating some of its computing capacity to internal AI development creates a challenge for the company as it seeks to meet customer requirements.
Pichai acknowledged that accommodating customers seeking additional computing resources for 2027 could be extremely difficult. He also explained that the infrastructure available today reflects investment decisions made years earlier.
Building new data centres and expanding computing capacity takes time. Companies must plan for hardware procurement, electricity supply, networking equipment, construction and other infrastructure requirements well before the resources become operational.
As a result, a sudden increase in demand cannot always be addressed immediately, even when a company has the financial resources to expand.
Why Computing Capacity Has Become a Major Issue
AI models require large amounts of computing power for training, testing and deployment. As models become more capable and AI services attract more users, the infrastructure required to support them also expands.
Technology companies are competing to secure advanced processors, memory, data centre space and reliable electricity supplies. Shortages in any of these areas can delay expansion and make additional computing resources more expensive.
Google faces two competing requirements. Its cloud business needs sufficient capacity to serve paying customers, while its AI researchers need powerful computing systems to develop and improve advanced models.
Pichai said Google plans its infrastructure requirements over a period of several years. However, the rapid pace of AI development makes it difficult to predict precisely how much capacity will be required in the future.
A model or service that attracts unexpectedly high demand can quickly place additional pressure on existing infrastructure. This means companies must continually adjust their investment plans and resource allocation decisions.
The Role of Gemini and Google DeepMind
Google's investment in frontier AI research is closely connected to its efforts to improve the Gemini family of models and develop new AI capabilities.
Pichai argued that advanced models provide an early view of emerging technological possibilities. This can help Google prepare future products and identify opportunities to improve existing services.
He also discussed the potential for AI models to assist researchers in developing subsequent generations of models. This process, often described as recursive self improvement, involves AI systems helping with activities such as coding, testing and research.
Such capabilities could eventually improve the efficiency of AI development. However, they do not eliminate the need for human oversight, research expertise and substantial computing infrastructure.
Google also continues to develop models designed for different requirements. Smaller models can offer faster responses and lower operating costs, while more advanced models may be better suited to demanding tasks.
Pichai indicated that Google intends to pursue a range of AI capabilities rather than rely exclusively on smaller models. The company believes maintaining a strong position in advanced AI research is important to its long term technology strategy.
Alphabet Plans Major Infrastructure Investment
Expanding computing capacity requires substantial capital investment. Alphabet, Google's parent company, has projected capital expenditure of 195 billion to 205 billion dollars for 2026, according to recent reports.
A significant portion of this investment is associated with AI and cloud infrastructure. Spending supports the expansion of data centres and the acquisition of equipment required to develop and operate AI services.
Nevertheless, additional investment does not immediately resolve capacity constraints. Construction timelines, equipment availability, power requirements and other practical limitations can delay the introduction of new infrastructure.
Google must also consider how demand may evolve over the next several years. Building too little capacity could restrict growth, while investing too aggressively could create financial risks if demand changes.
Pichai expressed confidence in the long term need for computing infrastructure while acknowledging that companies must make careful decisions about future requirements.
