Some Thoughts On NVIDIA Acquiring Arm

The big news to start the week is that NVIDIA has decided to acquire Arm. My first reaction is that it’s sad the company that still capitalizes the entire name has acquired the one who changed from that to normal capitalization. However, I decided not to end my analysis there…

Both companies are major players in the artificial intelligence (AI) space. NVIDIA began by providing GPUs to improve computer graphics, hence the name graphics processing unit. As last decade evolved, the company realized that deep learning could leverage the matrix build and software behind GPUs, and NVIDIA is the clear leader in the AI chip market. They, of course, subcontract manufacturing, but they directly market and sell their own products.

Arm took a different approach to technology. Rather than hardware, their focus has always been intellectual property (IP). It was founded in 1990 to create and license RISC architecture technology for chip manufacturing. It does not create its own chips but licenses the technology to whomever wishes to leverage it.

Both were successful models. Both companies also realized that AI was a growing sector and heavily invested in IP related to supporting the AI market. While Softbank purchased Arm in 2016, it’s clear they had no real vision for what to do with the company. Jensen Huang and his team at NVIDIA clearly realized that the company’s move to be more than a GPU provider, to continue to be a major player in the still nascent AI sector, meant that they could make far more use of Arm than could a diverse holding company.

The transaction, estimated at $40 billion, is aimed at extending NVIDIA’s reach in AI. While it has spent significant time and money on building the APIs needed to leverage GPU technologies higher in the development food chain, Arm already has a much wider breadth of software and market exposure.

While I can see there being the usual risks inherent in any large acquisition, there’s also the large upside of improving NVIDIA’s ability to remain in the lead in AI. Both sides, though, rely on Jensen Huang. He’s a brilliant leader, but he has sometimes changed focus a bit too quickly and without much planning. More importantly, I’m not seeing him letting go of much in the larger company. I’m reminded of Jerry Sanders, the legendary co-founder and long time CEO of AMD. He guided the company through phenomenal growth. He made some mistakes, but overall did a fantastic job. What he didn’t do was create a succession plan; and AMD has suffered from that since.

What will define the success of the merger and NVIDIA’s future is whether or not Jensen can expand his team, provide strong experience for future leaders, and create a succession plan that employees, investors, and other stakeholders can support.