Google is reportedly pushing more custom TPU adoption through cloud “neocloud” partners, aiming to pressure Nvidia as Wall Street trims big-tech expectations

Alphabet’s Google is reportedly trying to expand sales of its custom Tensor Processing Units, or TPUs, using a network of cloud providers that it characterizes as “neocloud” partners, according to a report cited in market coverage published Sunday.

The report, attributed to The Information and carried by Yahoo Finance, says Google is specifically exploring which neocloud providers it can work with to increase adoption of its custom TPUs, including Nvidia-backed Nscale. It frames the effort as part of an intensified push to compete in the infrastructure market for artificial intelligence workloads.

Nvidia, whose graphics processing units and related systems have become a default choice for many AI data center deployments, has spent the past two years selling not just chips but also a broader compute ecosystem, including software and interconnect technologies designed to support large AI training and inference clusters. Against that backdrop, any shift in how customers procure specialized AI accelerators can influence Nvidia’s pricing power and share in new deployments.

The market coverage also links the TPU push to a broader near-term sentiment shift. It notes “price target cuts” affecting big technology stocks ahead of the next round of earnings, suggesting investors are weighing not only product demand but also expectations for growth and margins in the sector.

Even with that competitive narrative, the details made public in the cited report remain limited. The market write-up does not lay out specific contract sizes, committed capacity, or the timing of any deployments. It also does not specify whether Google is offering custom TPU designs through these partners, bundling software, or changing procurement terms for customers.

For Nvidia, the relevance is twofold. First, custom TPUs are intended for AI workloads, competing directly with the accelerator role Nvidia’s offerings play inside data centers. Second, the mention of an Nvidia-backed provider implies that the competitive line may be less about “chip vs. chip” in isolation and more about how customers assemble and buy AI infrastructure from a mix of vendors, partners, and systems integrators.

What investors will likely watch next is whether Google’s channel strategy results in measurable customer adoption, whether competitors respond by changing partnership terms, and how analysts’ earnings models evolve. If the TPU expansion remains largely channel-based without adding disclosed new scale, market reaction may stay muted until customer or revenue disclosures provide clearer evidence.

As for what is still uncertain, the cited coverage does not provide further substantiation beyond the report’s claims, and it does not include quantified data. Google’s and Nvidia’s own official communications about TPU order books, partner pipeline, or competitive wins were not included in the information provided for this story. Any conclusion about revenue impact would therefore be speculative until companies make more explicit disclosures.