Google aggressively targets NVIDIA customers with TPU sales effort; signals major shift in AI chip competitive dynamics.

Google is reportedly expanding efforts to sell its custom tensor processing units (TPUs) to independent cloud providers as a direct challenge to Nvidia's dominance in AI hardware. According to The Information, the company has historically confined TPUs to its own facilities and rented them exclusively through Google Cloud, but is now taking its custom AI chips to market more broadly.

Google is specifically pitching TPUs to "neoclouds"—young companies built around renting out Nvidia GPUs. The report identified Nscale, a two-year-old neocloud where Nvidia is a major investor and preferred shareholder, as one of Google's targets. Google's pitch emphasizes more stable performance and simpler server networking compared to GPU-based systems, particularly as Nvidia's newer Grace Blackwell and upcoming Vera Rubin systems have caused deployment headaches for buyers.

Nvidia reportedly caught wind of the discussions and offered financial incentives to Nscale aimed at preventing the neocloud from adopting TPUs, though Nscale disputed this characterization. Beyond competitive positioning, the push for higher TPU adoption could also strengthen Google's hand in securing production capacity from Taiwan Semiconductor Manufacturing Company (TSMC), where Nvidia is currently the largest customer.

The reported TPU expansion arrives amid a broader reassessment of AI infrastructure spending on Wall Street. UBS on Monday cut its price target on Alphabet to $400 from $410, maintaining a Neutral rating, and lowered Meta's target to $766 from $865 while keeping a Buy rating—both moves tied to upcoming second-quarter earnings. Alphabet is scheduled to report Q2 results on July 22, with Wall Street expecting earnings of $2.90 per share on revenue of $116.8 billion.

Morgan Stanley, however, struck a more constructive long-term view, raising capital expenditure forecasts for Meta and Amazon. The bank now expects hyperscaler AI spending to reach roughly $1.2 trillion in 2027 and $1.4 trillion in 2028, reflecting continued investment in AI infrastructure despite near-term valuation concerns. Google's stock has gained over 10% this year and nearly doubled over the past 12 months.