Nvidia-Groq deal keeps fiction of competition alive: analyst says

It’s been two days since news broke that Nvidia was spending $20 billion to acquire top talent from Groq in what the chip startup called a “non-exclusive licensing agreement.”

Nvidia, the world’s most valuable company, hasn’t issued a press release or regulatory filing and, according to a spokesperson, is only confirming the contents of Groq’s 90-word blog post published after the close of holiday-shortened trading on Wednesday.

“They’re so big now that they can do a $20 billion deal on Christmas Eve with no press release and nobody bats an eye,” said Stacy Rasgon, an analyst at Bernstein, in a Friday interview with CNBC’s “Squawk on the Street.”

While neither company confirmed the price tag, CNBC learned from Groq lead investor Alex Davis on Wednesday that Nvidia had agreed to buy assets from Groq, a designer of high-performance artificial intelligence accelerator chips, for $20 billion in cash. Davis’ firm, Disruptive, has invested more than half a billion dollars in Groq and led the startup’s latest financing round in September at a $6.9 billion valuation.

Groq founder and CEO Jonathan Ross along with Sunny Madra, the company’s president, and other senior leaders “will join Nvidia to help advance and scale the licensed technology,” the startup said in the post, adding that it will continue as an “independent company,” led by finance chief Simon Edwards.

As an acquisition, Groq would mark by far Nvidia’s largest in its 32-year history. Its biggest prior purchase happened in 2019, when Nvidia bought Israeli chip designer Mellanox for close to $7 billion.

But Nvidia is instead following a playbook used by other tech giants over the last couple years, spending billions of dollars to hire top talent in AI and to get access to key technology through licensing agreements.

It’s a strategy that’s been employed by Meta, Google, Microsoft and Amazon. Nvidia itself has previously used the tactic, shelling out more than $900 million in September to hire Enfabrica CEO Rochan Sankar and other employees at the AI hardware startup, and to license the company’s technology, CNBC reported at the time.

By avoiding traditional acquisitions, tech companies have been able to skirt some level of antitrust scrutiny and quickly close deals to bring in the people they most covet.

“Antitrust would seem to be the primary risk here, though structuring the deal as a non-exclusive license may keep the fiction of competition alive,” Rasgon wrote in a Thursday note to clients. His firm recommends buying Nvidia shares and has a $275 price target on the stock.

Shares of Nvidia rose about 2% on Friday to $192.40. The stock is up 43% this year and is up thirteenfold since the end of 2022, when generative AI started taking off following the launch of OpenAI’s ChatGPT.

Nvidia has been using its expanding cash pile to invest capital across the AI ecosystem, including through recent investments in OpenAI and Intel. At the end of October, Nvidia had $60.6 billion in cash and short-term investments, up from $13.3 billion in early 2023.

Groq was founded in 2016 by a group of former engineers, including Ross. He was one of the creators of Google’s tensor processing units, or TPUs, the search giant’s custom chips that are being used by some companies as an alternative to Nvidia’s graphics processing units, or GPUs. 

The startup’s specialty is on the inference side of the market, which refers to the the use of AI to make decisions based on new information. Nvidia dominates the training piece of the market, which involves teaching AI models to learn from patterns in large amounts of data.

Analysts at Cantor said in a report Friday that Nvidia is “playing both offense and defense” by snapping up Groq’s assets, keeping them from potentially landing in the hands of a competitor.

“We think this acquisition only enhances Nvidia’s full system stack and overall leadership in the AI market (and only widens its competitive moat),” wrote the analysts, who kept their buy rating and $300 price target.

BofA Securities analysts also maintained their buy recommendation and $275 target following the announcement. In a note Friday, they characterized the deal as “surprising, expensive but strategic,” and said it shows Nvidia recognizes that “while GPU dominated AI training, the rapid shift towards inference could require more specialized chips.”

The analysts said that key questions remain, such as who will own Groq’s language processing unit intellectual property, whether it can be licensed to Nvidia competitors and whether what’s left of Groq — its nascent cloud business — could possibly “undercut NVDA’s LPU-based service with lower pricing.”

Nvidia isn’t commenting on any of those details for now. The first opportunity analysts and investors will likely get to hear from the company will be Jan. 5, when CEO Jensen Huang is scheduled to speak at CES in Las Vegas.

— CNBCs David Faber contributed to this report

WATCH: Nvidia GPUs to Google TPUs — Breaking down all the AI chips