Nvidia’s graphics cards fund the company’s entire empire, and that empire is now drawing the kind of regulatory attention usually reserved for railroads and phone monopolies. According to Bloomberg, the Justice Department has opened an investigation into whether Nvidia structured its $20 billion deal with AI chip startup Groq specifically to sidestep the antitrust review a conventional acquisition would have required.

The deal DOJ is examining

Nvidia and Groq announced the agreement on December 24, 2025, and it’s the largest deal Nvidia has ever signed. Rather than buying Groq outright, Nvidia agreed to pay roughly $20 billion for a non-exclusive license to Groq’s inference chip technology — the specialized hardware Groq uses to run trained AI models quickly and cheaply. As part of the same agreement, Groq’s founder and CEO Jonathan Ross, president Sunny Madra, and a number of other senior staff moved over to Nvidia to help build out the licensed technology.

Groq itself didn’t disappear. The company says it continues on as an independent business, now run by former finance chief Simon Edwards as CEO. On paper, nothing was acquired and no company changed hands.

A structure built to dodge merger review?

That’s precisely what regulators are questioning. Under the Hart-Scott-Rodino Act, a deal of this size — one company buying another — would ordinarily have to clear a federal antitrust review before it could close. License Groq’s technology and hire away its leadership team instead, and no merger filing is required at all, even though the practical effect looks a lot like an acquisition: Nvidia gets the technology and the people who built it, and a well-funded AI chip competitor is effectively sidelined.

Nvidia has defended the arrangement as a legitimate licensing partnership that accelerates AI inference development rather than a disguised takeover. The DOJ probe is still in its early stages and could conclude without any enforcement action.

Senators raised the alarm months earlier

The investigation follows scrutiny that started in Congress. In March, Senators Elizabeth Warren of Massachusetts and Richard Blumenthal of Connecticut sent a letter to Nvidia CEO Jensen Huang arguing the Groq deal appeared “to evade scrutiny by antitrust regulators,” and pressed the company for details on how the agreement was negotiated and structured. So-called “acqui-hire” deals — licensing a startup’s technology while absorbing its key employees, without a formal merger — have become increasingly common across the AI industry precisely because they sidestep traditional deal review, and lawmakers have been watching for a test case. Nvidia’s deal with Groq, given its size, became the obvious one.

Why gamers should care about an AI chip deal

Groq doesn’t make GeForce cards, so this isn’t a story about your next GPU upgrade directly. But it’s a window into how Nvidia is spending the enormous profits its gaming and data-center businesses generate, and how much leverage the company now has to reshape the AI hardware market without going through the regulatory process a normal acquisition would face. Nvidia has become the dominant supplier not just of gaming graphics cards but of the chips training and running much of the AI industry, and deals like this one extend that dominance further into AI inference hardware specifically.

There’s also a supply-chain angle gamers are already feeling. The same memory chips that go into GeForce cards — GDDR6X and its successors — are in tight supply partly because AI infrastructure spending is eating up fabrication capacity, a squeeze that’s already pushed up prices on other hardware this year. The more of the AI hardware stack Nvidia controls, the more its business priorities shape what gets built first and what gaming hardware waits behind it.

What happens next

The DOJ investigation is ongoing, and there’s no timeline yet for when regulators might decide whether to pursue formal action or close the inquiry. If the department does conclude the deal was structured to evade merger review, it could set a precedent affecting how every other “license and hire” AI deal gets structured going forward — including ones far smaller than Nvidia’s.

For more on how mega-deals are reshaping the games and tech business this year, see our coverage of EA’s $55 billion take-private buyout, and for the wider economic backdrop Nvidia’s dealmaking sits inside, our look at gaming’s record profits, record layoffs paradox.