Here’s How To Get Armed And Ready For The Hottest IPO In Ages

It’s been a while since we’ve seen any enthusiasm about an initial public offering (IPO). This year’s IPO market has been feeble, to say the least, and last year’s was no great shakes either. No surprises, then, that the public outing of Arm – one of the snazziest semiconductor companies around – has got people talking. I’ve been weighing it all up, and here are my thoughts about whether you should be jumping up and down about these freshly minted Arm shares.

What’s going on here?

Arm’s owner, the Japanese telecommunications firm SoftBank, is looking to offload 9.4% of its ownership in the British-based semiconductor firm by throwing some shares at the stock exchange. Retail investors won’t get a chance to invest in those freshly minted Arm shares until they start trading on the Nasdaq exchange though. That’s because IPOs are a kind of dog-and-pony show that take place between the firm's management, an army of investment bankers, and prospective investors – typically, the bankers’ biggest professional investor clients. The show kicked off this week, and a recent filing revealed a price range between $47 and $51 for the 95.5 million shares that’ll be on offer. All told, that’d value Arm somewhere between $48 billion and $52 billion.

After the hectic roadshow’s done, hopeful investors will place their “buy” orders and wait for the first trading day (looking like September 14th) to arrive. Sometime before the first trade, those professional investors will be told exactly how many of the 95.5 million shares on offer they’ve been allocated, and the price they’ll have to pay. And sure, the set range is $48-$51, but that could change, depending on demand. After that, all eyes will switch to the first day of trading to see whether the shares soar or flop. At that point, retail investors can decide whether to get involved. It’s all quite exciting, really.

Part of the reason why Arm’s IPO is drawing tons of attention is the fact that the market for brand-new listings has been dryer than the Black Rock Desert. Wait…

But, seriously, look at this chart. It shows the monthly value of US IPOs for the past few years. At $50 billion, Arm’s IPO would eclipse every listing since the end of 2021, combined.

US monthly IPOs by value, in billions of dollars. Source: Bloomberg.

SoftBank, then, is probably trying to take advantage of that IPO supply drought. It may also be hoping to be cute when it comes to timing. Let’s face it, unless you’ve been living under a rock, you know that semiconductors are at the center of everything important in the stock market these days – it’s basically all AI and Nvidia, all the time. In short, you’re going to want to take a front-row seat for the whole episode.

So let’s see what Arm’s all about – because this isn’t a regular semiconductor company.

What is Arm, then?

Well, it’s arguably the most important firm in the whole semiconductor ecosystem. The beating heart (well, brain, to be precise, but “beating brain” isn’t a saying) of all electronic devices is something called a central processing unit (CPU). These are chips that power smartphones, laptops, powerful server computers, and basically anything electronic.

Now, Arm doesn’t actually make these chips, the manufacturers do that – think: Intel, Texas Instruments, or TSMC. Arm spends its time developing the underlying architecture of a chip, and licensing that to other semiconductor companies that take the blueprints and add their own finishing touches. Think of Arm as the instruction kit for chips. Other firms take those instructions and add customizations before knocking them up, either by themselves (in the case of Intel or Texas Instruments), or with help from TSMC (in the case of Nvidia or Qualcomm, for example). Here’s Arm’s picture that captures all those words. (I’d be feeling pretty smug if there were 1,000 words. Don’t count – there are 147).

Here’s what Arm does, in a nutshell, as illustrated in its F1 filing with the Securities and Exchange Commission. Source: Arm’s SEC filing.

Because Arm has been doing this for a long time, and has unrivalled scale and expertise, it can create these instruction manuals at a fraction of what it would cost for firms to do it themselves. What that means is that more than 260 semiconductor companies ship products based on Arm’s technology, paying a tiny license fee for the blueprints. Arm estimates that more than 70% of the world’s population uses products based on its designs, which are found in 99% of the world's smartphones too.

Look, Arm’s got a hand in everything, so to speak. But here’s something: those license fees added up to only $2.7 billion in revenue for its most recent financial year (ended March 2023). Obviously, that’s not the sort of dough you find down the back of your sofa, but to be honest, stacked against Intel’s $50 billion, Texas Instruments’s $18 billion, or AMD’s $23 billion, Arm’s revenue number is surprisingly small, if you ask me. Mind you, that could be an opportunity, (more on that later).

What’s Arm’s story?

It hasn’t always been owned by SoftBank, of course. The firm bought Arm in 2016, and before that, it was jointly listed on the London Stock Exchange and the Nasdaq. (And, oof, the fact that SoftBank didn’t choose London for this latest listing is the source of much consternation among British politicians. But that’s a whole separate story).

The firm’s history dates back to 1990, when it was founded as a joint venture between Acorn Computers (older readers might nostalgically recall the Acorn Electron PC), Apple, and VSLI technologies (which later became NXP Semiconductors). Arm became a public company in 1998, around the middle of the dot-com boom, and went on to lavish tremendous riches on its lucky shareholders. Arm’s 1998 float price was a split-adjusted (meaning: taking into account all the share splits the firm has undertaken since) 43 pence ($0.40). SoftBank paid £17 per share for Arm in 2016, an eye-watering 3,853% gain for those initial stockholders.

Should you invest in the new listing, then?

Well, maybe, maybe not. (Sorry I can’t be more help on that.) The truth is, I genuinely don’t have a strong view on this myself. And I’m a tech investing nerd if ever there was one. However, I am following the whole thing closely, so I’ll share with you the pros and cons as I see them.

Pros:

1) It doesn’t matter that Arm’s blueprints aren’t currently being used to make the Nvidia graphics processing units (GPUs) that are front and center of the AI boom – demand for chips of all shapes and sizes is set to explode if AI computing really weaves its way into everyday life. The beauty of investing in Arm is that it means you don’t have to pick a winning chipmaker. The company’s designs are sold to more than 260 semiconductor firms. So it’s a broad-brush play on semiconductor demand.

2) Arm’s royalty payment business model is a smart one. It’s more repeatable and recurring than relying on the success of any one technology, and that makes the firm more valuable. In theory, I should be prepared to pay more for a dollar of Arm’s reliable, repeatable profit than a dollar of, say, Intel’s.

3) Remember I said that Arm takes a tiny fraction of a chip’s total manufacturing cost as a license payment? Well, if Arm is as essential to the industry as it says, there should be a long runway for raising prices. To me, that $2.7 billion revenue looks quite small, relative to the importance of the firm.

4) Arm could move more directly into AI computing chip designs. Nvidia can’t control everything, surely?

5) According to its F1 filing, Arm generated $671 million in profit last year, off $2.7 billion in revenue. That’s a 25% margin. Now, that’s not terrible, but given that license businesses are usually more profitable than that (Qualcomm, another semiconductor firm with a sizable licensing business made a 35% profit margin last year, and it used to do even better) I’d say there’s room for Arm to fatten its margin significantly.

Cons:

1) Arm isn’t squarely at the center of the AI revolution, so I do wonder whether it’s worth bothering with its stock at all. After all, I could just buy more Nvidia.

2) Arm’s price-setting bankers have been struggling to decide on a value. When SoftBank first said it was going to list some of Arm’s shares, the valuation was rumored to be around $60 billion. That then dropped to between $50 billion and $55 billion, and now it’s between $48 billion- and $52 billion. As any home seller knows, a surefire way to let people know that there mightn’t be too many buyers of your home is to publicly lower its price a few times.

3) Not long after announcing its intention to IPO 10% of Arm, SoftBank bought the 25% stake it didn’t already own from its technology investing fund Vision for a price that valued Arm at $64 billion. So either SoftBank and its bankers have no idea what Arm’s worth or Arm is suddenly worth a lot less than it was only a couple of months ago. That worries me.

4) Even at a reduced $50 billion, the price tag looks pretty steep to me. And I reached that conclusion by doing two things.

First, I compared Arm’s price-to-sales ratio (a popular valuation metric, which, at $50 billion, would be 19 times) and its price-to-profit ratio (75 times using the firm’s disclosed 2023 operating profit of $671 million) to a bunch of semiconductor companies. The chart below summarizes the multiple comparisons. With the exceptions of Nvidia and AMD (which has an elevated profit multiple because of a depressed level of profit), Arm is clearly priced high compared to its peers.

Price-to-sales and price-to-profit multiples for Arm and other semiconductor companies. Source: Koyfin.

Second, I used my favorite valuation technique for tech stocks – the reverse DCF. Effectively, this takes a valuation ($50 billion in Arm’s case), and a cash flow number, and with a bit of experimentation, it allows you to work out what cash flow growth rates the valuation is assuming. I wrote all about it here. Now Arm doesn’t disclose a cash flow figure, but I took a stab at estimating one by assuming the company would probably convert about 75% of its operating profit into cash – a reasonable cash-from-profit conversion rate for a firm like Arm. And the results from my valuation calculator tell me that Arm would need to grow cash flow by around 16% for the next 10 years and then 2.5% after that. That’s not inconceivable for an AI-powered firm, but remember that Arm isn’t quite that, yet, so that forecast could be a tad on the ambitious side.

Arm’s reverse DCF shows that a $50 billion valuation assumes a 16% rate of cash flow growth for 10 years and a 2.5% rate thereafter. Source: Arm’s SEC filing.

Of course, valuation is just one piece of the puzzle. If you’re a believer that all semiconductor demand is set to explode and that Arm is, in fact, nicely positioned to reap some AI rewards, then it may be an attractive proposition, even with its reasonably expensive IPO price. What I would say, is that the share price is likely to be volatile, and whether it skyrockets or plummets in the days after the IPO will tell you a lot about how the broader investment community feels about the firm’s prospects. For me, then, letting the IPO dust settle might just be the way to go, before jumping in and making a rash trade.