Nvidia's Groq Arrangement Is Now a Federal Question — and It Lands While the Company Is Betting Big on Three Continents

DOJ reviews whether Nvidia's $17B Groq licensing deal dodged antitrust scrutiny, as the chipmaker pushes a $12.93B Hugging Face acquisition and global buildout.

Nvidia's expansion playbook is running on multiple tracks at once, and Washington has just inserted itself into the middle of it. The U.S. Department of Justice is examining whether the chipmaker deliberately structured its licensing arrangement with AI startup Groq to sidestep antitrust review, according to a New York Times report picked up by Reuters. A formal request for information has reportedly been delivered to the company.

The timing could hardly be more loaded. Nvidia is simultaneously pushing through its roughly $12.93 billion acquisition of Hugging Face — one of the largest deals in its history — which Reuters says is expected to close in the first half of 2027, subject to regulatory approval. The more scrutiny lands on how Nvidia conducts its business, the more uncertain that timeline becomes.

What's Actually at Stake

The question investors need to answer isn't whether the Groq matter is uncomfortable. It's whether it stays contained or becomes a template. If the inquiry hardens into a broader examination of Nvidia's licensing agreements, infrastructure partnerships and acquisitions, the company's entire growth architecture — from Australia to Southeast Asia to the Hugging Face deal — moves under the antitrust microscope. That distinction is the whole ballgame.

The Groq arrangement itself is unusual by design. Rather than acquiring the company outright, Nvidia secured non-exclusive technology rights and brought key personnel into the fold, including founder Jonathan Ross. Critics have a name for this pattern: a "reverse talent acquisition." The arrangement is valued at $17 billion. Should regulators object, the likely outcome is fines rather than an unwinding — but that doesn't make the regulatory climate any less of a wildcard.

The Expansion Case Is Real

Against that backdrop, Nvidia's operational momentum hasn't slowed. Together with partners Firmus, CDC, NEXTDC and AirTrunk, the company is planning a sweeping buildout of data center capacity in Australia — up to two gigawatts of AI capacity by 2027. In Southeast Asia, expansion via the Zankore platform, developed with Indosat Ooredoo Hutchison and Nokia, is being financed through a credit facility of up to $3.1 billion, according to Reuters.

The consumer flank is moving too. Nvidia's collaboration with Lenovo and Acer will put the first Windows PCs running the RTX Spark chip on shelves in October.

On the Goldman Sachs conference stage Thursday, CEO Jensen Huang reaffirmed guidance of roughly 70% revenue growth for fiscal 2027 and identified cybersecurity as the next major AI application area, backed by partnerships with CrowdStrike, Cisco and Palantir. Nvidia is also running Palantir's Foundry and Nemotron technology through its own supply chain — a signal that the company stress-tests its AI tools internally before selling them into manufacturing, energy, healthcare and aviation.

From Chips to Collateral

Huang's more consequential argument is that Nvidia's systems should no longer be understood as depreciating hardware but as durable collateral — what Morgan Stanley has dubbed "Balance Sheet as a Service." The bank raised its price target to $300 while staying cautious on credit risk. By the end of 2028, Morgan Stanley projects roughly $200 billion in credit exposure, including about $170 billion in contingent liabilities from leases, residual value guarantees and revenue participations.

That framing invites the obvious criticism. The Bank for International Settlements warned in its annual report that circular AI financing represents a trillion-dollar systemic risk, and investors are asking whether Nvidia's stakes in CoreWeave, OpenAI and other customers amount to buying demand that later gets presented as organic growth. Huang rejects the charge outright: the investments are "immaterial" relative to the business they generate, he says, and the company only enters "sure-win" deals. That's persuasive as long as underlying demand holds — but it doesn't shrink the balance sheet risk on its own.

Where the Stock Sits

The market has been pricing in both sides of this story. Nvidia shares currently trade at EUR 189.44, just 2.4% above the 50-day moving average but 6.4% below the 52-week high of EUR 202.50. Another reading puts the stock at EUR 189.82, up 0.9% on the prior close and about 6.3% below its May 2026 peak. Either way, the picture is the same: sideways to slightly weaker, despite a 25% gain over twelve months. Thirty-day annualized volatility of 40% tells you the market is caught between the growth narrative and a risk premium it can't quite shake.

Piper Sandler and Morgan Stanley have both set $300 targets, citing the Rubin architecture's promise of thirty times the throughput per megawatt versus Grace Blackwell Ultra, with falling token costs. That's the technological case — strong, but not beyond debate.

Two Catalysts to Watch

For now, as long as the Groq inquiry remains an information request without a formal complaint, the expansion story — Australia, Southeast Asia, Hugging Face — likely stays the dominant narrative. If that changes, whether through widening investigations or early regulatory objections to the Hugging Face acquisition, the risk premium on Nvidia shares should widen with it.

Two concrete markers lie ahead: the October launch of RTX Spark PCs, and the progress of the Hugging Face approval process, targeted for the first half of 2027. Investors would do well to track both threads — operational expansion and regulatory exposure — because together they will determine where the stock goes next.