A familiar problem is haunting Nvidia. The chip maker's stock has paused in its latest effort to break out to new highs and part of the problem is Google.

Nvidia shares were down 0.6% at $227.47 in premarket trading on Wednesday. The stock is having a tough time surpassing its record closing high of $235.74, reached on May 14.

Part of the issue is just the same macroeconomic considerations hitting the wider chip sector-the PHLX Semiconductor Index is down 5.9% over the past three months. But Nvidia constantly has to fend off concerns that competitors are eating into its dominant artificial-intelligence chip share.

The biggest threat is Alphabet's Google and its Tensor Processing Units, designed in collaboration with Broadcom. Although Nvidia and Google's rivalry is generally an amicable one, investors keep a close eye on which chips the biggest players use.

So it could be disquieting to hear that Anthropic is talking about leasing up to one gigawatt in compute capacity from Stream Data Centers, largely powered by TPUs, according to technology-focused news outlet The Information. Anthropic and Stream Data Centers didn't immediately respond to requests for comment from Barron's early Wednesday.

That comes a week after a syndicate of 10 banks agreed to provide a $22 billion loan to Crux AI, a cloud-computing venture backed by Blackstone and Alphabet, providing funds for purchasing TPUs.

It isn't necessarily a big problem for Nvidia. After all, Anthropic recently signed a cloud-computing deal worth $35 billion with cloud provider Lambda that will use Nvidia chips, and a $45 billion deal with another Nvidia-backed neocloud, Nscale, according to The Wall Street Journal.

But it's a reminder that big clients such as Anthropic are constantly trying to reduce their dependence on Nvidia. In August, Anthropic confirmed it was putting together an in-house chips team, following in the footsteps of ChatGPT-developer OpenAI.

As Barron's has previously written, Nvidia is fighting back by backing the open-source AI community to boost a wider range of customers. But it takes a lot of small AI companies to replace one Anthropic and that's the pressure the company is constantly battling.