Lee: Is Nvidia Grace a sign of doubt?

Nvidia AI supercomputing Grace CPU
Nvidia Grace is an upcoming data center CPU, while Nvidia is mainly known for its GPUs. Analyst Leonard Lee wonders if Nvidia is getting ready in case its acqusition of Arm fails to get regulatory approval. (Nvidia)
Leonard Lee

During his opening keynote at GTC21, Nvidia CEO Jensen Huang announced Project Grace, a CPU based on Arm’s Neoverse geared for the data center. The company will be joining the ranks of AWS, Google and Qualcomm who license Arm’s CPU cores or ISA in pursuit of the emerging AI opportunity in the data center.

Nvidia’s entry into the data center CPU game has been widely celebrated as a landmark move that poses an existential threat to Intel’s dominance in the data center. Accordingly, Nvidia’s stock jumped 5.62 percent upon the announcement while Intel’s dropped 4.18 percent. However, the advent of Project Grace feels like something else. It could suggest that Nvidia sees the writing on the wall and needs a hedge in the event that Nvidia’s $40 billion acquisition of Arm falls through.

Given recent complaints filed by some of Arm’s marquee licensees Microsoft, Google and Qualcomm earlier this year, the Nvidia-Arm acquisition is now facing at least a year of regulatory limbo. That’s a year that Nvidia cannot afford to delay its ambitions against the backdrop of an Intel commandeered by new CEO Pat Gelsinger hellbent on restoring the company to its former glory and technical leadership.

In retrospect, Huang gave us a hint that a hedge was in play during his February 25th interview on CNBC with Jim Cramer who asked if Nvidia’s long-term growth would be negatively impacted if its acquisition of Arm was blocked. Huang stated, “Nvidia is going to be huge no matter what.” In other words, Nvidia doesn’t need to acquire Arm to achieve its goals and continue to grow its business in the data center and edge computing markets. Project Grace is effectively a statement of proof. Technically, the idea of acquiring Arm has been relegated to the status of nice-to-have by none other than the Nvidia CEO himself.

These latest remarks and actions bear stark contrast to Jensen’s sentiments expressed back in September 2020 when the Nvidia acquisition of Arm was first announced. Jensen Huang stated that the combination of Nvidia and Arm would create the “world’s premier computing company for the age of AI.”

However, the announcement of the deal drew early antitrust and geopolitical concerns as well as a sharp rebuke from Hermann Hauser, a co-founder of Acorn Computers who helped pioneer the Arm architecture. Dr. Hauser asserted in his letter to the U.K. House of Commons foreign affairs committee that the union would create “a quasi-monopoly supplier of microprocessors to the world.” Dr. Hauser also warned that the deal would end Arm’s position as the “Switzerland of the semiconductor industry.”

In response, Huang insisted that a “firewall” between the two businesses would be established and maintained to ensure that anticompetitive dynamics could not play out to the detriment of Arm’s customers and Nvidia’s competitors in the Arm ecosystem.

Regardless, in February of this year, the FTC issued a second request and opened up a probe into Nvidia’s acquisition of Arm upon complaints issues by several of Arm’s marquee customers and partners such as Microsoft, Qualcomm and Google.

Facing what will undoubtedly be a protracted FTC investigation, the timing of Nvidia’s announcement of Project Grace is curious. Project Grace could undermine Nvidia’s case to acquire Arm. It highlights the anticompetitive threat that Nvidia can present if it begins to directly compete with Arm customers and ecosystem partners who are eyeing the enterprise data center and emerging edge cloud market.

Project Grace also brings into question Nvidia’s rationale and intent for acquiring Arm. Is it to acquire Arm’s customer base? Is it end market share acquisition? Is it control of the direction of the Arm technology evolution or ecosystem?

While Nvidia has been undergoing a reinvention of its business, at its core, it is still primarily a GPU company. It’s hardware, software and solutions are GPU-centric. For this reason, it is difficult to see how Nvidia will advance Arm’s CPU-centric technology portfolio as a combined concern. Arm already partners with Nvidia and many of Nvidia’s competitors in the GPU space. In many ways, Arm is already well positioned to support its customers in becoming competitive computing companies in the era of AI, including Nvidia.

The elephant in the room is the real risk that Nvidia would pose in a combined company to influence the Arm technical road map in a way that provides Nvidia with a competitive advantage against Arm customers competing in the same markets.

It’s difficult to imagine any “firewall” governance model that could materially mitigate any undue influence that Nvidia could have on Arm without an exorbitant regime of monitoring and enforcement by global regulatory bodies including the FTC.

Of course these efforts would come at great cost to taxpayers and the industry at large. In light of a new Congress and a Biden Administration with sights set on reforming antitrust laws and tightening regulations, it is likely that the FTC will try to avoid repeating the buyer’s remorse situations that it currently suffers with certain members of the “Big Tech” crew.

At best, Project Grace proves that Nvidia is and will continue to be a great customer and partner of Arm in the era of AI. At worst, Project Grace may be the first self-inflicted nail in the coffin of the Nvidia+Arm deal.

Leonard Lee is the founder and managing director of neXt Curve, a research advisory firm focused on Information and Communication industry and technology research. He has worked as an executive consultant and industry analyst at Gartner, IBM, PwC and EY and has advised leading companies globally on competitive strategy, product and service innovation and business transformation. Follow Leonard on LinkedIn: linkedin.com/in/leonard-lee-nextcurve