The question is the structure, not the price

Merger control is triggered by the shape of a transaction rather than by its effect. That is a deliberate design choice, and it creates a standing incentive to produce the effect in a shape the trigger does not recognise.

Nvidia shares fell $5.41, or 2.42%, to $218.01 by mid-morning on Thursday after Bloomberg reported that the Justice Department is reviewing whether the company's $20 billion arrangement with the AI chip startup Groq was structured specifically to limit antitrust scrutiny, as relayed by International Business Times.

The arrangement, first unveiled in December, gave Nvidia access to Groq's underlying chip technology while leaving Groq operating as a separate, independently funded company rather than being absorbed. Groq's chief executive, Jonathan Ross, and its chief operating officer, Sunny Madra, subsequently moved to Nvidia.

Set out in sequence, the components describe something familiar. A buyer obtains the technology, the founders and the operating leadership. What it does not obtain is the corporate entity, and the corporate entity is what a merger filing is about.

Nvidia's position is that the agreement lets it use Groq's technology across its product range while Groq's cloud business continues to be run and funded independently of Nvidia's control. That is a coherent description of the paperwork, and the paperwork is the thing under examination.

Lawmakers have questioned whether the arrangement effectively amounts to a full acquisition, one they argue could reduce competition in AI chips without ever having faced the review a conventional merger would attract.

The pattern is not new and is not confined to semiconductors. Licence the technology, hire the team, leave the shell funded and independent: it has appeared repeatedly in AI over two years, and each instance has been defended on the ground that the target still exists.

Whether the target still exists is the wrong question. The question competition authorities have been circling is whether a company that has lost its technology, its founders and its operating leadership remains a competitor in any sense that matters to the market.

An inquiry is not a case. Reviewing how a transaction was structured is ordinary agency work and frequently ends without action, and the reporting describes a review rather than a complaint.

The market reaction was proportionate to that distinction. A 2.42% move on an inquiry into a deal already nine months old prices some probability of an adverse outcome, not an expectation of one.

What makes this one worth following is precedential rather than financial. If the agencies establish that a licence-plus-talent structure can be examined as an acquisition, the structure loses the property that made it attractive, and $20 billion of engineering around the filing threshold stops working.