The Justice Department is formally investigating whether Nvidia structured its $20 billion deal with AI chip startup Groq as a licensing agreement to avoid the mandatory antitrust review that a traditional acquisition would have triggered. On December 24, 2025, Nvidia announced it would pay $20 billion for a non-exclusive license to Groq's inference technology and hire away Groq's founder and president, Jonathan Ross, along with president Sunny Madra. No stock changed hands. No board seat was taken. Now, the DOJ's Antitrust Division wants to know if this arrangement was designed to look like something other than what it effectively was: a takeover.

What Makes This Deal Structure So Controversial?

The Hart-Scott-Rodino Act requires companies to notify federal antitrust regulators before closing any merger or acquisition valued above $119.5 million, giving the DOJ or Federal Trade Commission a window to object before the deal closes. Nvidia's Groq arrangement was valued at roughly 167 times that threshold, yet it involved no stock purchase and no merger filing. The question regulators are asking is straightforward: can a company avoid mandatory review by licensing technology and hiring key employees instead of buying shares ?

Groq makes a chip called the LPU, or Language Processing Unit, built specifically for AI inference, the fast, repeated work of running an already-trained model. It represents one of a handful of real challengers to Nvidia's dominance in that segment. The payment structure moved quickly: roughly 85 percent of the money to Groq shareholders upfront, another 10 percent by mid-2026, and the remainder by year end. A CNBC analyst summed up the market's interpretation on announcement day, saying the arrangement let Nvidia absorb a competitor's talent and technology while keeping just enough distance between the two companies to preserve, in the analyst's words, "the fiction of competition".

Why Are Lawmakers and Regulators Concerned?

Three months before the DOJ escalated its review, Senators Elizabeth Warren of Massachusetts and Richard Blumenthal of Connecticut sent Nvidia CEO Jensen Huang a letter laying out the same theory in plain language. The senators argued that by licensing Groq's technology and hiring its most important employees, Nvidia had effectively acquired Groq in all but name.

The senators' letter leaned on Nvidia's own market dominance to make the foreclosure case. Nvidia controls an estimated 81 to 90 percent of the AI-accelerator market, depending on methodology. The company's data-center segment generated $193.7 billion in fiscal 2026, up 68 percent year over year. A company that dominant, the letter argues, does not need to buy a rival outright to neutralize it; it can simply absorb the people and the patents and let the shell keep operating as cover.

"By licensing its technology and hiring its most important employees, NVIDIA has effectively acquired Groq in all but name," stated Senators Elizabeth Warren and Richard Blumenthal in their March 2026 letter to Nvidia CEO Jensen Huang.

Senators Elizabeth Warren and Richard Blumenthal, U.S. Senate

How Does This Deal Compare to Nvidia's Other Recent Acquisitions?

Nvidia's subsequent actions have only sharpened the contrast. In August 2026, just months after the Groq arrangement, Nvidia agreed to buy AI-model platform Hugging Face outright for $12.9 billion, a straightforward stock acquisition that, unlike the Groq arrangement, cannot avoid Hart-Scott-Rodino review. Analysts have noted the contrast directly: Nvidia's Groq and Poolside deals were both licensing-and-hiring arrangements that stayed under the antitrust radar; the Hugging Face deal is a conventional purchase that has to clear regulatory scrutiny. Nvidia disputes that the pattern reflects any intent to evade scrutiny.

"The Groq story is a prime example of the American system working as designed to promote innovation, reward entrepreneurs, and benefit consumers," stated an Nvidia spokesperson in September 2026.

Nvidia Spokesperson, Nvidia

What Are the Key Differences Between the Groq Deal and Traditional Acquisitions?

• Ownership Structure: The Groq deal involved no stock purchase and no transfer of board seats, whereas a traditional acquisition would require both and trigger mandatory antitrust review under Hart-Scott-Rodino.
• Regulatory Threshold: Nvidia's $20 billion Groq arrangement was valued at 167 times the $119.5 million threshold that triggers mandatory federal review, yet avoided that review entirely through its licensing-plus-hiring structure.
• Practical Control: The deal gave Nvidia access to Groq's core inference technology through a non-exclusive license and placed Groq's leadership directly on Nvidia's payroll, effectively consolidating control without formal acquisition.
• Contrast with Hugging Face: Nvidia's August 2026 acquisition of Hugging Face for $12.9 billion was structured as a direct stock purchase, triggering the antitrust review that the Groq deal avoided.

What Happens Next in the Investigation?

Nothing about the DOJ's current posture forces a resolution soon. A formal request for information is not a civil investigative demand, and neither is a filed complaint; it is the stage at which a company is asked to produce documents while regulators decide whether there is a case at all. Nvidia's stock dipped roughly 1 to 2 percent on the news, a modest move for a company still trading near its 52-week high.

The larger fact is procedural, not financial: the agency that blessed the AI boom's biggest supplier is now the same agency deciding whether that supplier's newest deal was designed to stay one step ahead of the law written to catch exactly this kind of transaction. The DOJ's Antitrust Division has jurisdiction over this matter because of a June 2024 agency-turf agreement that split AI-sector oversight, with the FTC taking Microsoft and OpenAI's conduct and the DOJ taking Nvidia. This probe sits inside a broader Nvidia inquiry the DOJ opened in August 2024, which produced subpoenas the following month.

The investigation reflects a fundamental tension in tech regulation: as companies grow more dominant, the tools available to challenge them through traditional merger review become less effective if those companies can achieve the same practical outcome through alternative deal structures. Whether the DOJ can successfully argue that Nvidia's arrangement should have triggered the same scrutiny as a stock purchase remains an open question, one that could reshape how regulators approach future technology deals.

" }