2. What’s the pitch?
Arm aims to reduce its reliance on a stagnating smartphone market by pursuing customers that build chips for data centers. Because it has expertise in the energy-constrained world of battery-powered phones, Arm argues that its products are a better fit for data centers that consume immense amounts of electricity when running the latest generation of artificial intelligence platforms. Among public companies, the main beneficiary of the current AI boom has been Nvidia Corp., which built an early lead for so-called AI accelerators — chips that excel at crunching data essential to the artificial intelligence training process. Giddy investors pushed Nvidia’s market capitalization above US$1 trillion for the first time in May. On Aug. 8, the company unveiled an updated AI “superchip” that boosts its flagship product’s capacity and speed. Half of that two-semiconductor package, the microprocessor, is built on Arm’s technology.
3. What companies rely on Arm?
Amazon.com Inc., Samsung Electronics Co. and Apple Inc. are also among Arm’s most important customers. Arm’s instruction set is in billions of devices, and the effort required to switch to another company’s code would be enormous. The main concentration is in the processors that run virtually all of the 1.4 billion smartphones sold every year, with more than 99% using either Arm’s designs or the instruction set. More recently, major tech names such as Amazon have designed their processors for data centers with those blueprints. That’s taking Arm into the most profitable part of the semiconductor industry.
4. Why is it so strategically important?
Arm has a relatively small workforce of about 6,000. But few companies reach so far across the tech ecosystem: More than 250 billion chips have been made with Arm technology. The company estimates that soon, 100% of the world’s digital data will be processed by that technology at some point during its life cycle. That makes Arm a virtual industry standard, meaning that everyone from software engineers to chip designers and electronic-device producers know they’re not going to have to duplicate their efforts if they switch suppliers. Its strategic importance is so great that when SoftBank decided in 2020 to sell the company to Nvidia, it prompted an outcry from Arm’s customers that eventually killed the US$40 billion deal. SoftBank’s Plan B was to sell Arm shares in New York in an IPO.
See also: NTT DC REIT's IPO 4.6 times subscribed
5. Why was the Nvidia deal scrapped?
Critics said a takeover would threaten a cornerstone of Arm’s success: its neutrality. Arm has been used across the US$550 billion semiconductor industry on the understanding that no one would get privileged access to its technology. The transaction unraveled after the US Federal Trade Commission sued to block it, and Nvidia walked away in February 2022.
6. So what is SoftBank’s plan?
SoftBank is expected to list a minority stake in Arm in New York. A bid for a dual listing in London failed when the government was unable to give the company a waiver on rules that would have made it difficult for SoftBank to do transactions without seeking shareholder approval. Despite Arm’s British roots, SoftBank dropped the idea, dealing a blow to a UK government that had lobbied hard to land the deal. SoftBank’s takeover of Arm came with promises that the Japanese company would create more jobs in the UK and leave the headquarters where it was.
7. What’s Arm valuation?
When SoftBank announced plans to take Arm public last year, the company was considering a valuation of about US$60 billion for the business. Later, when chip valuations fell, SoftBank decided to sell a smaller portion of Arm than previously planned and retain a controlling stake in the hope of obtaining a higher valuation for the remainder later. Since then, the benchmark Philadelphia Stock Exchange Semiconductor index has rallied again, fueled by the interest in AI. That’s given Arm a better environment in which to sell shares. But there’s still concern that the industry may be making too many chips. Those conflicting currents have made it hard to put reliable valuations on semiconductor companies.
