NVIDIA Corporation (NASDAQ:NVDA)’s $20 billion deal for Groq’s AI technology is facing a new legal challenge just as that technology begins contributing to Nvidia’s push into inference.
A lawsuit filed by two former Groq engineers in Delaware alleges that Groq’s board improperly structured the 2025 transaction, which consisted of a $17 billion licensing agreement and a separate $3 billion Nvidia stock bonus pool for certain engineers who joined the chipmaker. The plaintiffs argue that the arrangement effectively transferred Groq’s core technology and much of its engineering talent without treating the transaction as a conventional acquisition. Groq has called the lawsuit meritless.
For Nvidia investors, however, the bigger question is not simply whether the transaction faces litigation. It is whether Groq has become important enough to Nvidia’s next phase of growth for that legal and regulatory uncertainty to matter. NVIDIA Corporation also made our latest list of the 10 best AI stocks to buy in October, but another AI stock ranked even higher.
Photo by Javier Esteban on Unsplash
Groq Is Already Moving From Licensing Deal to Product
That distinction matters because Groq is no longer a speculative technology bet.
NVIDIA Corporation said Groq 3 LPX, its first rack-scale LPU system combining Nvidia’s high-throughput architecture with Groq’s high-interactivity technology, is already in full production, with volume shipments expected later this quarter. Nvidia said the system achieved 3,400 output tokens per second in an Artificial Analysis benchmark, delivering four times faster responsiveness than the nearest alternative platform.
Groq technology is also part of Nvidia’s broader Vera Rubin platform. Management estimates that its revenue opportunity per gigawatt has expanded from roughly $18 billion with Hopper to $25 billion with Blackwell and $40 billion with Vera Rubin, which now spans GPUs, CPUs, networking, and Groq LPUs.
However, Jensen Huang has indicated that Groq accelerators primarily target specialized, high-speed inference workloads, while Vera Rubin NVL72 remains the company’s broader AI training and inference platform.
That makes the lawsuit more relevant than the $20 billion headline alone indicates. Nvidia is trying to defend its dominance as AI spending increasingly moves from training toward inference, where specialized architectures and custom silicon represent a more meaningful competitive threat.
That competitive shift also puts Nvidia’s position against Broadcom under greater scrutiny, particularly as investors weigh which AI chip stock is better positioned as spending expands beyond traditional GPUs.
Piper Sandler captured that opportunity in September when it initiated NVIDIA Corporation at Overweight with a $300 price target, describing the company as the “enterprise inference king.”
Nvidia’s Valuation Already Assumes Enormous Earnings Growth
The financial backdrop makes the Groq question easier to frame.
Nvidia generated $96 billion of revenue in fiscal Q2 2027, more than double the prior-year level, while Data Center revenue reached $89 billion. Management expects approximately 70% revenue growth in fiscal 2028 and says that outlook is supply-constrained.
Morgan Stanley sees potential upside even to that forecast. The firm argued on October 2 that underlying demand could support growth closer to 100%, while Nvidia’s broad customer base and ability to generate more tokens per gigawatt leave it well positioned as power and data-center capacity become the next AI bottlenecks. Morgan Stanley has an Overweight rating and $300 target.
Those expectations explain why NVIDIA Corporation’s valuation looks very different depending on how far forward investors look.
The same valuation question becomes even more pronounced when Nvidia is compared with another megacap technology leader, raising the question of whether NVDA or Apple offers the stronger valuation case at current expectations.
NVIDIA Corporation traded at a 24.88 times forward P/E as of October 6. Consensus estimates call for EPS growth of 95.23% in fiscal 2027 and another 69.66% in fiscal 2028, bringing estimated P/E down from 25.65 times fiscal 2027 earnings to just 15.12 times fiscal 2028 earnings.
Put simply, Nvidia does not necessarily look expensive if those earnings materialize. The risk is that its valuation increasingly depends on maintaining extraordinary growth as inference becomes a larger part of the AI market.
The Bigger Risk May Extend Beyond This Lawsuit
For now, the Groq litigation does not appear large enough on its own to challenge Nvidia’s earnings trajectory. But the dispute raises a broader issue.
The plaintiffs acknowledge that Delaware courts have not directly determined whether acqui-hire transactions like Groq’s should face traditional M&A scrutiny. Separately, the U.S. Department of Justice was reportedly probing Nvidia’s Groq transaction, while FTC Chair Andrew Ferguson has said the agency would examine deal structures potentially designed to escape conventional merger review.
That matters because Nvidia did not merely obtain Groq technology. It also hired much of the engineering team behind it without buying the entire company. Greater scrutiny of that structure could affect how Nvidia and other technology companies acquire strategically important AI capabilities in the future.
Before the lawsuit emerged, hedge fund interest in Nvidia had increased. The number of hedge funds bullish on NVIDIA Corporation increased from 275 to 285 in Q2, with Fisher Asset Management increasing its position 3%, AQR Capital Management 18%, and Arrowstreet Capital 10%.
The near-term test, therefore, is not the lawsuit itself. It is whether Groq 3 LPX scales commercially and helps Nvidia extend its dominance from AI training into inference. If it does, the $20 billion transaction becomes easier to justify against the earnings growth Nvidia is expected to deliver. If legal or regulatory scrutiny begins interfering with that commercialization, the Groq dispute becomes much more relevant to NVDA’s valuation.
READ NEXT: Can Paramount (PSKY) Turn Warner Bros. (WBD) Scale Into Returns After its Regulatory Concessions? and Netflix (NFLX)’s Sports Bet Just Got More Specific. Is that Enough to Move the Needle?
