Nvidia will acquire chip designer Arm Limited from SoftBank Group for $40 billion in cash and stock, the companies announced late Sunday.
The mega-deal had been rumored for months and is expected to take 18 months to close, following regulatory reviews by the U.S, the U.K., China and the European Union.
Some analysts and existing customers of both companies had recently expressed concerns over how the combined entity would function and if it could bind-up deals and customer competitiveness. Major semiconductor companies could be concerned that Nvidia will have access to Arm IP before the general market gains access, critics have said.
But Jensen Huang, founder and CEO of Nvidia, assured customers that Arm will continue to operate its open-licensing model and will maintain customer neutrality.
Arm will also remain headquartered in Cambridge. Arm and Nvidia will also build the world’s largest supercomputer there as part of an AI research center, relying on both ARM CPUs and Nvidia GPUs.
Arm is “most passionate about openness and the neutral way they meet customers,” Huang said. “This is good for Nvidia. Arm has a vast network and we can help with accelerated computing and AI. We retain incredible benefits by protecting the Arm business model. Customers will see the logic in it.”
Huang noted that Arm’s chip designs are vast, with Arm having built designs for 180 billion computers over three decades and 22 billion last year alone. By contrast, he noted in a call with reporters that Nvidia built far fewer computers, 100 million, last year adding that Nvidia is in a “select set of markets.”
The combination of Nvidia GPUs and Arm CPUs will help Nvidia grow, he said. “We need to grow all over the world. This is a great fit for us,” Huang added. Nvidia, the ninth biggest semi maker in 2019 with $9.5 billion in revenues, is already an Arm customer, as are Apple, Qualcomm, AMD, Intel and many others.
In response to a question about how flexible Nvidia will remain under the Arm acquisition, Huang said customers will be able to buy Nvidia chips outright, but also could license Nvidia/Arm cores or modify Nvidia chips to build their own server CPUs. In a coming world of thousands of distributed data centers, “it’s not possible for one company to build it all,” he added. “We’ll have an entire network of partners around Arm and semi-custom or have a chip we made. We’d like the ecosystem to be as rich as possible.”
Arm CEO Simon Segars joined Huang in welcoming the deal as a step forward into advancing artificial intelligence at the edge, saying the two companies share a vision for “ubiquitous, energy-efficient computing to help address the world’s most pressing issues from climate change to healthcare and from agriculture to education.”
SoftBank will remain a 10% shareholder in Nvidia, allowing SoftBank to continue to invest in Arm’s success, SoftBank CEO Masayoshi Son, said in a statement. SoftBank bought Arm for $32 billion four years ago, and some terms of that deal will remain in effect for another year.
Nvidia will pay SoftBank $21.5 billion in Nvidia common stock and $12 billion in cash, including $2 billion. SoftBank may receive up to $5 billion in cash or stock, subject to specific financial performance targets by Arm. In addition, Nvidia will issue $1.5 billion in equity to Arm employees.
Huang said regulators in the various countries will see that the Arm and Nvidia combination is complementary and protects market competition. “We don’t license IP to semiconductor companies and we don’t participate in the cell phone market like Arm,” he added.
The combined entity will be able to accelerate the development of server CPUs. “Customers will be thrilled,” Huang said.