Arm, a British Chip Designer, Juggles Challenges Before 2023’s Biggest I.P.O.

The SoftBank-owned company, which for decades has defined how mobile phones operate, is grappling with geopolitical and commercial complexities as it prepares to go public this week.
Arm, a British Chip Designer, Juggles Challenges Before 2023’s Biggest I.P.O.

Rene Haas, the chief executive of the chip-design powerhouse Arm, has many masters to serve.

He reports to Masayoshi Son, the head of SoftBank, which owns Arm and plans to sell a portion of the British company this week in the year’s biggest initial public offering. Officials in Beijing and Washington also command Mr. Haas’s attention amid a widening chip trade war, as does Britain’s prime minister, Rishi Sunak, and others who have unsuccessfully pitched the idea of a stock listing in the country.

And Mr. Haas must juggle the demands of more than 200 companies that use Arm’s technology. Ten of the biggest — including Apple, Google, Samsung and Nvidia — have been negotiating for stakes in the highly anticipated Arm offering as artificial intelligence drives explosive demand for more powerful chips.

“It’s only going to get more and more complex,” Mr. Haas said in a speech in May at a trade show in Taiwan. “I’m an old person in this industry. I’ve never seen it like this.”

Few companies face as many geopolitical and commercial complexities as Arm, the creator of the most widely used computing architecture of all time. Its public offering, which is expected to start trading on Thursday and to value the company around $52 billion, will signal Arm’s ability to weather those challenges and enter new markets. How Arm performs will also influence the market for public listings, which has been quiet for much of the year.

There are “a lot of companies to please and a lot of capitals to please,” Jodi Shelton, chief executive of the Global Semiconductor Alliance, a big trade group, said of Arm. Mr. Haas, she added, “has to not only play diplomat on a global level but to customers.”

Arm is in a quiet period before its public offering, with Mr. Haas expected to help ring the Nasdaq opening bell in New York on Thursday.

Founded in 1990, Arm has for decades helped define how nearly all mobile phones operate, despite having only about 6,000 employees and less than $3 billion in annual revenue. Its technology has also spread to cars, sensors, supercomputers and myriad other devices.

Unlike most chip companies, Arm does not make or sell those silicon components. It essentially designs and licenses blueprints for one of the most important parts of a chip — processor cores, which perform calculations and run software programs.

Those electronic brains rely on a set of instructions that Arm developed, and that programs like Google’s Android and Apple’s iOS operating systems use to carry out basic operations on smartphones.

Companies typically design entire chips by placing the blueprints for Arm’s cores alongside those handling other functions, which they may design or license. That means many companies watch Arm’s technical specifications and future plans closely.

“There are very different interests that have to be balanced,” said Aart de Geus, the chief executive of Synopsys, which sells software for designing chips along with specialized chip cores and aims to invest in Arm’s offering. “We all have to play really well together to make anything work.”

Arm estimates that more than 250 billion chips using its technology have been sold since its founding in 1990. That was when Apple and two partners established what was originally called Advanced RISC Machines in Cambridge, England. Apple wanted low-powered technology that would extend battery life in the Newton, its ill-fated personal digital assistant, which was discontinued in 1998.

“The DNA of the company was born building products that run off batteries, and that sensibility still defines us today,” Mr. Haas said in a video presentation for Arm’s I.P.O. roadshow, where it pitches to investors.

Mr. Haas, 61, is the first American to run Arm, which still conducts most engineering in Cambridge. He grew up in a suburb of Rochester, N.Y., where his father was a Xerox research scientist and exposed him to Silicon Valley’s allure on a childhood visit to Xerox’s research center there.

After earning an electrical and electronics engineering degree from Clarkson University, Mr. Haas worked for a chip maker and a start-up before a seven-year stint at Nvidia, a dominant provider of graphics and artificial-intelligence chips.

He joined Arm in 2013. After lobbying for changes in its China business, Mr. Haas was transferred to Shanghai.

In 2016, SoftBank bought Arm for $32 billion, inspired partly by Mr. Son’s ultimately unsuccessful idea of selling services to help coordinate and deliver software to billions of devices equipped with Arm chips. Mr. Haas moved to London to lead the rest of Arm’s business, and began pushing for changes there.

Among other things, he introduced licensing schemes with annual subscriptions for a bundle of Arm technologies, reducing the need for repeated negotiations for individual products. Arm also shifted from a longtime practice of designing cores for smartphones and adapting them for other applications. It now designs cores from scratch for new markets like data centers and cars.

In September 2020, Nvidia reached a deal to buy Arm from SoftBank for $40 billion. That plan collapsed 18 months later after opposition from regulators and customers. Mr. Son picked Mr. Haas in February 2022 to succeed Simon Segars as chief executive.

An imposing figure at 6 feet 4 inches, Mr. Haas introduced new ideas with a collaborative style, associates said. He “has really transformed Arm,” said Jensen Huang, Nvidia’s chief executive, in the roadshow video.

With Arm in the middle of tech supply chains and the chip trade war, Mr. Haas is grappling with challenges, including slowing sales of smartphones and the question of whether the company can play a big role in A.I. chores in data centers.

Another comes from China, where Arm gets roughly a quarter of its revenue and licensing is handled by Arm China, a company that it does not control. Arm weathered a long-running battle with Arm China’s leader, who was ousted last year, but payments and sales information from that company have sometimes been late, the I.P.O. prospectus said.

Trade tensions also loom large. Sales of one powerful version of an Arm core for data centers have been hobbled by American and British restrictions on Chinese exports. Though Arm said it had worked around such limits so far, the possibility of tougher rules stands out among numerous risks concerning China described in the prospectus.

There are also questions about Arm’s business model. Arm historically negotiated with chip makers to receive an upfront licensing fee, which might range from $10 million to $100 million depending on the technology involved, plus a per-chip royalty of around $1 to $3.50, estimated Jim McGregor, an analyst at Tirias Research.

But lower-cost alternatives have emerged, including RISC-V, a technology whose instruction set can be licensed for free.

Arm also faces a legal battle with a major customer, Qualcomm, a big supplier of mobile chips. Arm sued Qualcomm, saying it violated a licensing contract in connection with buying a chip start-up, Nuvia. Qualcomm denied the accusations and plans to begin offering chips developed by the former Nuvia team, underscoring how more customers may license just the Arm instruction set and then design original processor cores.

Though Arm has penetrated data centers in chips designed by Amazon and Ampere Computing, a start-up, it has yet to experience the kind of A.I. sales surge there that Nvidia has enjoyed.

Handel Jones, head of the research firm International Business Strategies, said Arm’s biggest A.I. opportunity would be in phones, PCs and other so-called edge applications where devices need low power consumption.

“If Arm can play the role that Nvidia plays in data centers at the edge, then that I.P.O. valuation can be justified,” Mr. Jones said. “They are a long way from getting there yet.”