As investors flock to AI stocks, Softbank is trying to market Arm as an AI company.
Following an extended drought in tech listings, the impending IPO of Arm is the talk of the town in Silicon Valley and Wall Street these days. Ever since the sale of the UK-based chip designer to Nvidia fell through, tech executives and investors alike have been eager to find out how the linchpin of the mobile semiconductor realm intends to proceed. But now that the moment is here, the hotly anticipated IPO is stirring up quite a bit of controversy.
The bone of contention is Arm’s valuation. Softbank, which took Arm private in 2016, is looking north of 60 billion dollars. That’s a hefty price for a company that would be worth closer to half that when applying the industry-average price-to-sales multiple. What could justify that kind of premium?

Lofty
The recent financial results aren’t very convincing. Arm reported a slight revenue decline last year as consumers are holding on to their phones for longer, illustrating the company’s heavy reliance on the smartphone market. It’s great to be a monopolist, but when there’s no more market share to conquer, results can only move with the ebb and flow of consumer spending in a market that’s left its high-growth years behind it.
There are, of course, plenty of prospects for growth in other markets. But these tend to be much more competitive environments, where Arm has yet to prove it can cut out a substantial slice of the pie for itself. It’s not at all like the hay days of the smartphone market, which was essentially driven by a single end-customer: Apple. Over the past years, Arm has successfully made inroads into faster-growing markets such as the IoT and embedded, only to find these aren’t nearly as profitable as mobile. Furthermore, in these relatively low-margin markets, the royalty-free RISC-V architecture is an appealing alternative, especially for new entrants that have to compete with large incumbents that have leverage to negotiate with Arm over license fees.
Larger Arm licensees, too, are embracing RISC-V as a way to avoid becoming overly dependent. The prospect of Arm being acquired by Nvidia served as “a wake-up call” to the industry, said NXP CTO Lars Reger. Along with four other European chipmakers, NXP recently started a RISC-V company, which will initially focus on automotive, but will expand into mobile and IoT as well.
By far the largest threat to Softbank’s lofty valuation is Arm China. For reasons that still aren’t entirely clear, Softbank sold a 51 percent share in its Chinese subsidiary to state-connected investors in 2018 and subsequently lost control over the company altogether. As a result, there’s no guarantee that revenues from China – about a quarter of the total – will continue to make it back to Cambridge.
Track record
The reason Softbank thinks Arm is still worth 60+ billion dollars is artificial intelligence. Expected to be one of the major drivers of semiconductor growth in the next decade, sales of AI chips are booming even right now, during a slump in the semiconductor market. Softbank contends that Arm has a lot to contribute to the AI revolution and should therefore be valued more as an AI company like Nvidia, the poster child of AI success, and less than Qualcomm, a smartphone chip designer – even when Arm has yet to show similar growth prospects to the GPU maker.
There are still plenty of opportunities for Arm to grow beyond smartphones, be it in PCs, notebooks or servers, in cars or, in time, in AI. No doubt the British company will present some compelling business cases in an upcoming roadshow ahead of the IPO. However, with sluggish results, a competing architecture gaining traction, a dissident Chinese subsidiary and no proven track record in the industry’s biggest growth market, Softbank is probably overplaying its hand.
Update 6 September: Softbank has reportedly lowered Arm’s price tag to 55 billion dollars