Alphabet’s expanding artificial-intelligence chip business is central to a bullish case from J. Morgan analyst Doug Anmuth, who recently lifted his price target for the company’s stock to $425. That target represented about 23% upside from Alphabet’s share price at the time of the assessment.
Alphabet has designed Tensor Processing Units for its own data centers for years and has recently begun offering them to selected outside customers. Chief Executive Sundar Pichai said the company plans to deliver TPUs for use in customers’ data centers, including applications involving AI laboratories, capital-markets firms and high-performance computing. Custom processors can be designed for particular models and workloads, potentially making them more efficient for some AI tasks than relying entirely on general-purpose GPUs.
The opportunity is emerging alongside rapid growth in Alphabet’s cloud business.
Anmuth’s argument assumes that cloud growth can accelerate as Alphabet sells more AI infrastructure, including its own processors. TrendForce estimates that custom-processor sales will grow 45% over the following year, compared with 16% growth for GPU sales, while Piper Sandler analyst Thomas Champion has estimated that Alphabet’s processor sales could reach $104 billion by 2028. That figure is a projection, not reported revenue.
Alphabet is also reporting heavy demand across its wider AI infrastructure portfolio. The company says its models are processing about 22 billion API tokens per minute, up from 16 billion one quarter earlier, and that more than 9 million developers use its models each month. It has identified demand from AI laboratories, Kakao, Deutsche Börse, Pfizer, Roche and World Labs, while its new Axion CPU is described as delivering 30% better performance per dollar than peer offerings.
The bullish case faces a significant spending test. Alphabet expects to invest heavily in AI infrastructure, and investors remain focused on whether revenue from cloud services and future TPU sales can justify the cost of data centers, servers and networking capacity. The company is scheduled to report its third-quarter results on Oct. 28, giving investors the next major update on cloud growth, AI demand and the early financial impact of its external-chip strategy.