Nvidia-Groq Deal Draws Antitrust Heat In AI Chip Race

Nvidia’s dominance in AI chips is now drawing attention not only because of what it sells, but because of how it grows around the rest of the market.

US regulators are reportedly investigating Nvidia’s licensing arrangement with Groq, the AI chip company known for its language processing units and fast inference systems. Reports say the Department of Justice is looking at whether the deal functioned like an acquisition while avoiding the normal scrutiny attached to one.

The structure is what makes the story interesting. Nvidia did not simply buy Groq outright. The reported arrangement gave Nvidia access to Groq technology through a large non-exclusive licensing deal while senior Groq executives, including founder Jonathan Ross and COO Sunny Madra, moved to Nvidia. Coverage of the probe says regulators are now asking whether that kind of deal can reduce competition without triggering a formal merger review.

That is a serious question in the AI chip market. Nvidia already sits at the centre of AI training, inference, networking, software and cloud relationships. If rivals lose key technology or leadership through licensing and hiring arrangements, competition can weaken even when the rival company technically remains independent.

Lawmakers had already raised this concern months ago. Senators Elizabeth Warren and Richard Blumenthal publicly questioned whether Nvidia’s Groq deal was structured to avoid antitrust law. Their argument was that AI companies should not be able to buy the practical value of a competitor without going through the same checks that would apply to an acquisition.

For Nvidia, the counter-argument is simple. Licensing is common in technology, talent moves happen across the industry and Groq can still operate as a company. The problem is that AI has made every major deal look strategic. A chip design, a software stack, a cloud contract or a team of engineers can all shape who controls the next phase of computing.

The timing also matters because Nvidia has become the clearest financial winner of the AI boom. We recently wrote about Nvidia’s massive AI-driven quarter, and the numbers explain why regulators are looking more carefully. The bigger Nvidia becomes, the more every adjacent deal will be read through a competition lens.

Groq is important because inference is becoming the next battlefield. Training frontier models gets the headlines, but running those models cheaply, quickly and at scale is where the daily business of AI will happen. If Groq’s technology gives Nvidia more leverage in inference, regulators will want to know what that means for buyers and rivals.

This does not mean the deal will be unwound. These investigations often end with document demands, fines or new conditions rather than dramatic reversals. But the signal is clear enough: AI deal-making is entering a tougher regulatory era.

The lesson for the industry is that soft acquisitions may no longer stay soft. When a dominant AI company uses licensing, hiring and investment to absorb strategic value from a rival, regulators may treat the substance of the deal as more important than the label on the paperwork.