The Arm IPO: A Look Under the Hood

Just over six years, we wrote a piece titled How to Invest in the Singularity – It’s Near which dissected a talk being given by Masayoshi Son, CEO and Founder of SoftBank. The thesis was quite simple. SoftBank acquired Arm Holdings in 2016 for $32 billion because they believed it would be a critical provider of information for the AI algorithms of tomorrow. At that time, Mr. Son expected to ship 1 trillion IoT chips in the next 20 years with Arm IoT chips commanding an 80% market share. That would allow SoftBank to plan the future direction of IoT which would also influence the direction of AI significantly.

SoftBank’s convictions may have been called into question when they later decided to sell Arm Holding to NVIDIA. Our excitement was short-lived when the deal fell through last year, and here we are looking at an Arm Holdings F-1 statement as the company prepares for the biggest IPO in nearly two years.

Arm’s Fab-u-less Business Model

We’ll typically wait for an IPO to take place before investigating a company, because there’s no guarantee an F-1 filing leads to an IPO. In this case we’d like to look at what’s under the hood before a potentially hyped IPO. At an expected valuation of between $60 billion and $70 billion, Arm would have a simple valuation ratio (SVR) of 22 based on the lower end of that guidance. That’s just above our cutoff of 20, and already showing signs of exuberance. But with a gross margin of (checks notes again) 96% last year, and market leadership across multiple domains, it’s understandable why everyone’s getting excited.

To understand how any company might achieve such high gross margins we need to examine their business model. Arm describes theirs succinctly as follows:

Arm licenses processor designs to semiconductor companies that incorporate the technology into their computer chips. Licensees pay an up-front fee to gain access to our technology and a royalty on every chip that uses one of our technology designs. Typically, the royalty is based on the selling price of the chip. An Arm design can be used in many different chips, and certain Arm designs have shipped for more than 25 years.

Arm F-1

The below visual shows just how effective this business model is in case the 96% profit margins weren’t evidence enough.

ARM Business model
Credit: AnandTech

Arm is a fabless semiconductor company, which means they don’t actually build chips. They don’t ship chips either, their clients do. Developing chip designs, licensing them to partners, then collecting downstream royalties is how you manage to achieve incredibly high gross margins of 96%. Arm seems to have the Holy Grail of fabless semiconductor business models which customers across multiple industries can’t get enough of.

Chips Eat the World

Over 99% of the world’s smartphones use Arm CPUs which power everything from the tiniest of sensors to the most powerful supercomputers. The company breaks down their business activities into a number of areas along with accompanying market share for each.

  • IoT and embedded chip market – grew from 58.4% in 2020 to 64.5% as in 2022
  • Networking equipment market – grew from 18.8% in 2020 to 25.5% in 2022.
  • Cloud computing market – grew from 7.2% in 2020 to 10.1% in 2022
  • HPC systems, enterprise servers, and edge networking equipment – grew from 9.1% in 2020 to 16.2% in 2022
  • Consumer electronics – around 32% in 2022

Chips are found in nearly everything we use, and Arm commands nearly 50% market share for CPUs globally. As Mr. Son quipped, during some time in the next 30 years, the two chips in each of your shoes will be more than twice as smart as you are – and you will step on them every day.

Fit for Royalty

A breakdown of revenues coming from each segment would be insightful, but instead, we’re provided with a categorization that includes “Licensing” and “Royalties,” the latter of which accounted for 59% of last quarter’s $675 million in revenues.

Bar chart showing ARM's Quarterly Revenues

Royalty revenues typically have no associated cost of goods sold which means they represent pure profit. They’re also consistent because customers are contractually obligated to pay them. Even the predictable revenues of a SaaS firm can be blindsided by a cancelation, but royalties keep on accruing as long as end customers are buying products. The below chart shows how resilient Arm-powered chip shipments were when The Rona unleashed havoc upon the planet.

Bar chart showing number in billions of Arm-based chips shipped
Credit: Arm Holdings

Given Arm’s domination in smartphones it’s tempting to think that the five largest customers of the company which account for nearly 60% of revenues just might be smartphone makers. Could their second-largest customer which accounts for 11% of total revenues be Apple? We can’t say for sure, but we’re certain that Arm does sell lots of chips to the world’s largest smartphone makers. Geographical revenue breakdowns also tell us something about where Arm’s customers are, and particular attention should be paid to the “PRC” category (25% of total revenues) as that could get caught up in the ongoing geopolitical tension between China and the USA.

Arm's geographical revenue segmentation
Credit: Arm S-1

“…we expect to continue to see declining royalty revenues derived from the PRC and we could see a decline in licensing revenues.”

Arm F-1

Customer concentration risk is less of a concern for Arm given their business model which drives 60% of revenues from downstream royalties. Their product is so pervasive it’s hard to see a single large customer suddenly deciding not to work with them. Even if that happened, they could still realize royalties for decades to come. But China, that’s a different story, because it represents geopolitical uncertainties that can’t be controlled. Monitoring China will be easy as it’s also counted as a single customer and Arm’s largest.

Would We Invest in the Arm IPO?

Investing in a volatile IPO is never a good idea, so our tech investing methodology calls for a waiting period of one month for newly listed stocks. Let’s set aside any concerns this IPO might not happen and assume the simple valuation ratio eventually settles below our cutoff of 20. The first concern would be an overexposure to semiconductors with NVIDIA now occupying over 12% of our current portfolio (guess we’ll begin trimming again to take that down to 10%). As long as NVIDIA continues growing, we have no need to sell the stock. So let’s pretend NVIDIA isn’t an issue.

We’re particularly interested in the various segments that Arm has laid out with accompanying market share. Certain disruptive tech themes drove growth for Arm in 2021 such as 5G smartphones which use more Arm technology than 4G smartphones. The ongoing global rollout of the 5G network has driven demand for Arm chips which are used in base stations and wireless networking equipment. The data centers of some cloud service providers also utilize Arm chips, and providing more granular revenue segmentation would help us understand what’s growing or lagging at any given time.

A 96% gross profit is a cash printing machine, so we’d like to understand more about what competitive threats they’re facing in each segment. After the IPO, the business should provide key metrics for investors to monitor. With 80% of the company’s employees focused on research, design, and technical innovation, they’re clearly focused on maintaining their competitive position and moving towards Mr. Son’s vision of a trillion chips by 2030. Even if they fail at that endeavor, royalties will sustain the company for years to come.

…approximately 46% of our royalty revenue for the fiscal year ended March 31, 2023 came from products released between 1990 to 2012.

Arm F-1


The next step is to hurry up and wait for the IPO debut. Once the dust has settled, we’ll see how the valuation looks and what additional insights the company offers in investor decks. Should we ever move out of our NVIDIA position, Arm might make for a compelling replacement at the right price. We’re adding Arm Holdings to our tech stock catalog as a like and will check back after the IPO dust has settled to see what valuations are sitting at.


Leave a Reply

Your email address will not be published.

    1. That’s certainly something we’ll consider in a future article. Unfortunately we have a word limit on these pieces and can’t cover everything. The article specifically says, “we’d like to understand more about what competitive threats they’re facing in each segment” so thank you for raising that!

    1. Aside from noting that shares are extremely overpriced (SVR of over 40), this article is what we think of ARM right now. We’ll likely check back in with the company over the summer. Hopefully the AI hype will have subsided by then.

    1. You’re a Premium annual subscriber so we’re going to move your request to the front of the queue. This is a relevant and interesting company to cover (most likely in a video).