Two former Groq engineers have filed a proposed class action in Delaware's Court of Chancery against Groq's former board and ex-CEO Jonathan Ross, alleging a $20 billion deal handed Nvidia the LPU technology and the team that built it while keeping billions from other shareholders.

The suit, filed by Benjamin Serebrin and Joshua Rubin and unsealed Monday, challenges a structure that combined a $17 billion license fee for all backers with a separate $3 billion Nvidia stock pool reserved for engineers who transferred. The plaintiffs concede that "no Delaware decision has directly answered the question this case raises," but the deal already faces a reported Department of Justice inquiry.

What does Nvidia's own accounting reveal?

Nvidia's annual report (10-K) books $14.4 billion of the license as goodwill "primarily attributable to the workforce and future development of the licensed technology," plus $2.5 billion for the technology itself. The goodwill figure covers roughly 85% of the $16.9 billion Nvidia booked — about 5.8x the value assigned to the technology asset.

The 10-K notes state that "no customer contracts, existing products, or equity interests were purchased," with "$13.0 billion paid at closing and $4 billion, inclusive of imputed interest, payable within one year." In short, Nvidia placed far more value on the LPU design team and future development than on the silicon itself.

What hardware changed hands?

The deal covered both the LPU and its creators. The LP30 packs 500MB of SRAM per die with 150 TB/s of bandwidth, uses Samsung's 4nm process, and Nvidia rates it at up to 1.23 FP8 PFLOPS. A single LPX rack holds 256 LPUs delivering 128GB of SRAM, 40 PB/s of bandwidth, and 315 FP8 PFLOPS in aggregate, operating as a decode co-processor for Vera Rubin NVL72.

The LP30 appears to match the next-gen LPU Groq roadmapped in August 2023 on Samsung's SF4X process with its first-generation Tensor Streaming architecture. Ross, a former Google TPU engineer, founded Groq in 2016.

At Hot Chips in August, Igor Arsovski — Groq's former chief architect, now Nvidia VP of hardware — called the event "a pinch me moment for the Groq team that's now integrated into the Nvidia group." He said the rack was already in production. That supports the complaint's factual premise, though not the legal question of whether more is owed.

What do the plaintiffs allege?

The suit claims the board was conflicted, failed to secure the best terms for every shareholder, and denied some of them a vote. Ross and other top employees were allowed to "take a discount on those shares and be paid separately for following the technology to Nvidia." Nvidia hired "nearly all" of Groq's engineers — as many as 200.

Additional allegations include:

• The $17 billion license payment was taxable as income at Groq because Nvidia did not buy the company outright.
• The Series A valued the remaining Groq above the price paid to shareholders bought out.
• Four "conflicted funds" on the board — BlackRock, Social Capital, Infinitum, and Disruptive, not named as defendants — engineered that gap and profited by staying after the buyout.
• The final payout ignored potential upside and synergies with Nvidia.

Axios reported in December 2025 that sources close to the deal said most shareholders would get per-share payouts "tied to the $20 billion valuation." Around 90% of employees would join Nvidia, with vested shares paid in cash and unvested shares converted to vesting Nvidia stock. Groq told CNBC the suit is "meritless" and said the licensing agreement "delivered exceptional value for Groq, our investors, and our employees."

Where does Groq stand now?

Groq said at announcement it would continue "to operate as an independent company." In August it joined the Nvidia Cloud Partner program, closed a $350 million Series A at a $3.5 billion valuation, and said it is "among the first adopters of NVIDIA Groq 3 LPX" with Vera Rubin NVL72, deploying with Dell. Before the deal, in September 2025, its valuation stood at $6.9 billion after a $750 million round.

Groq now operates 13 data centers and expects to scale from 54 to 200+ MW in 2027. The Financial Times reported Groq "pivoted entirely to AI cloud computing, dropping its chip design efforts." Groq claims to be "the only team with hands-on experience operating LPUs in production at scale."

What are the regulatory implications?

The New York Times reported in September that the DOJ sent Nvidia a formal request for information, with a fine possible. Nvidia called the Groq story "a prime example of the American system working as designed." Senators labeled it a "reverse acqui-hire" in February, and FTC Chair Andrew Ferguson said in January the agency is "beginning to examine these acqui-hires to make sure they are not an attempt to get around" merger review.

When the deal was announced, Nvidia CEO Jensen Huang told employees in an email obtained by CNBC that while Nvidia was adding Groq staff and licensing its IP, "we are not acquiring Groq as a company." The suit's outcome carries implications for similar structures, including Nvidia's pending Hugging Face agreement — an outright purchase, but one that also splits purchase price from employee equity.

For LPX customers, Nvidia names Nebius the first AI cloud to adopt the racks, online later this year. Its roadmap continues with the LP35 (NVFP4 support), Rubin Ultra, and the LP40 for Feynman.