SoftBank plans Arm IPO instead
Nvidia’s two-year effort to acquire semiconductor maker Arm will end in defeat, according to a Bloomberg report. Nvidia first announced plans to acquire Arm from SoftBank Group in a $40 billion deal in September of 2020. Japan’s SoftBank acquired Arm in 2016 for $32 billion.
Nvidia almost immediately encountered industry and regulatory resistance. Hyperscalers and other semiconductor makers like Intel and Qualcomm sounded alarms. Nvidia offered EU regulators concessions in October. Then in December, the U.S. Federal Trade Commission (FTC) sued to block the acquisition. That would seem to have been the knockout blow for the troubled acquisition.
FTC Bureau of Competition Director Holly Vedova said the FTC made its decision to help preserve a competitive and healthy semiconductor economy.
“This proposed deal would distort Arm’s incentives in chip markets and allow the combined firm to unfairly undermine Nvidia’s rivals. The FTC’s lawsuit should send a strong signal that we will act aggressively to protect our critical infrastructure markets from illegal vertical mergers that have far-reaching and damaging effects on future innovations,” said Vedova.
Bloomberg reports that Nvidia has told partners it doesn’t expect the deal to close. It also said that SoftBank is stepping up plans for an Arm public stock offering instead. Bloomberg cites multiple anonymous sources to corroborate the report. Nvidia and Arm offered Bloomberg diplomatic non-answers in response to questions about the deal.
According to the original signing deal, SoftBank retains $2 billion from Nvidia if the deal terminates unsuccessfully.
Nvidia’s data center ambitions
Nvidia’s ambition to acquire Arm hasn’t quelled its efforts to get its hardware out as far as possible. Nvidia has risen from just a performance GPUs maker for gamers and 3D modelers. Now the company is a formidable fabless semiconductor company that wields global influence. Nvidia ranks only behind Qualcomm and Broadcom for aggregate semiconductor revenue for 2020, according to a Trendforce report.
Nvidia’s data center revenue has been setting records. It’s found new markets for its GPUs in cloud and data center hardware. Nvidia now makes specialized high-performance compute (HPC) systems, specialized systems made for Artificial Intelligence (AI) and Machine Learning (ML), and accelerated IT workloads. The company’s Omniverse, its metaverse project attracted Ericsson. Ericsson and Nvidia are using it to create digital twin cities. The goal is to configure optimal placement for 5G network buildouts prior to actual construction. Nvidia’s Omniverse recently emerged from beta.
Nvidia’s DGX AI platform is the compute technology powering Meta’s recently-announced AI Research SuperCluster (RSC). The RSC is a supercomputer the rebranded Facebook is using to build new AI models. Meta wants to use its modeling to help produce safer online experiences and, ultimately, rich metaverse interactions, it said in a statement.
