Google Prioritizes AI Over Cloud Clients

Google is deliberately foregoing short-term revenue from its cloud business to prioritize the development of advanced AI models, according to CEO Sundar Pichai. This strategic decision means that limited computing capacity is being allocated to internal AI research, such as that conducted by Google DeepMind, rather than fulfilling all demands from paying cloud customers.

Pichai emphasized that securing compute power for frontier AI development is "non-negotiable." He explained that this focus allows Google to gain early insights into new AI abilities, which is crucial for preparing future products. The company anticipates that these advanced models will also contribute to their own improvement through recursive self-assistance.

The tech giant's cloud customers are experiencing challenges in securing the necessary computing resources due to a broader industry-wide shortage of chips, servers, and data-center space. Pichai noted that accommodating new customers for capacity in 2027 would be "incredibly difficult," highlighting that current compute availability is a result of long-term planning decisions made years in advance.

Alphabet is investing significantly in capital projects, with an expected expenditure of $195 billion to $205 billion in 2026, primarily to bolster AI and cloud infrastructure. Despite a substantial contracted revenue backlog for Google Cloud, estimated at $513.9 billion, the demand for computing power, even from Google's own AI products like Gemini 4 Argon, is stretching existing capacity.