- Market cap of some $50 billion looks lofty to annual sales of $2.6 billion and annual net income of $524 million
- Surge in the Philadelphia Semiconductor Index (itself fired by NVIDIA) provides a relatively helpful benchmark
- Investors may be wary of large offering syndicate and Softbank’s prior attempt to set a valuation through a prior transaction
- Limited free float could help squeeze the shares higher
- Advisers and bankers will be hoping a successful deal opens the door for further primary offerings and listings in Q4
“The price range set by Softbank and its army of bankers and advisers for ARM of $47 to $51 a share implies a total market capitalisation for the business of $47 billion to $51 billion (since it has a billion shares in issue) with its $2 billion net cash pile on top. That gives an enterprise value (EV), or total value for the business of $49 billion to $53 billion, and it is now the job of investors to decide whether that is fair price or not, and one that gives buyers a chance of making a fair risk-adjusted return over time,” says AJ Bell investment director Russ Mould. “This can be done looking at ARM’s history and the valuation of quoted peers. Softbank’s decision to list the firm in New York may have the latter methodology in mind but neither approach suggests ARM is returning to the public markets after seven years at a knock-down price. The share price in the after-market will now tell us whether moving ARM’s listing to New York from London is the correct thing to do – or not.
“Based on the company’s first-quarter results, ARM will do well to generate flat sales and profits in its fiscal year to March 2024, and that comes after a flat year in 2023. The F1 regulatory filing released last week revealed that between April and June 2023, ARM generated $675 million in sales, 2% lower than in the same period in 2022, and net profit fell by more than half to $105 million from $225 million.
|
$ million |
Fiscal year to March |
|
First quarter to June |
|||
|
|
2021 |
2022 |
2023 |
|
2022-23 |
2023-24 |
|
Sales |
2,027 |
2,703 |
2,679 |
|
692 |
675 |
|
Operating profit |
239 |
633 |
671 |
|
294 |
111 |
|
Operating margin |
11.8% |
23.4% |
25.0% |
|
42.5% |
16.4% |
|
Net profit |
388 |
549 |
524 |
|
225 |
105 |
|
EPS ($) |
$0.53 |
$0.66 |
$0.51 |
|
$0.22 |
$0.10 |
Source: Company accounts. 2021 numbers impacted by discontinued operations, Treasure Data and IoTP.
“This is in keeping with forecasts of a difficult overall backdrop for the semiconductor industry. According to industry specialists such as WSTS, SIA and Gartner, global silicon chip sales will drop by some 10-12% in 2023 before rebounding by a similar degree in 2024. Hopes for a second-half recovery in 2023 have been dented somewhat by profit warnings from the likes of Analog Devices and Micron, although NVIDIA’s upgrade to third quarter estimates after its demolition of consensus for the second quarter offers some hope.
Source: WSTS, SIA, Gartner
“Overall, consensus analysts’ forecasts believe that total sales for the 30-stock Philadelphia Semiconductor Index, known as the SOX, will drop by just 2% to $462 billion in 2023, before a 14% increase in 2024. Net profit is seen dropping 15% before a 40% surge in 2024.
“It is therefore upon the potential for a 2024 rebound that hopes lie for the SOX and, by implication, ARM, especially as the SOX index is up by 45% this year and sits just short of its all-time highs.
Source: Refinitiv data
“However, the SOX includes nine semiconductor production equipment companies, and it may be best to exclude those, as their business models bear no resemblance to that of ARM. That said, ARM is a very special silicon chip business in that, historically, its approach has been both fabless and chipless – it does not manufacture semiconductors in factories (or fabs) or even provide its own designs.
“Rather, the company has focused on providing core architectures upon which customers can then build their own designs more rapidly and cost effectively. ARM charged a licence fee for its architectures and then banked a royalty every time a product was sold that featured a chip designed using one of those architectures. There are, however, suggestions that ARM has started to look into developing its own chips beyond prototype level, although the firm has never confirmed or denied this.
“A valuation benchmark can be derived, to some degree, from looking at the 21 silicon chip firms in the SOX, however some of those are fabless, some are chipless, some are foundries (third-party manufacturers), and some make their own product. Again, therefore, the comparison is not direct.
“Thanks to NVIDIA’s weighty contribution, the SOX trades on market cap-to-sales multiples of nearly 10 times and price-to-earnings 44 times and 31 times multiples of for 2023 and 2024 respectively.
“ARM is going to be priced at a premium to that.
Source: Company accounts, Marketscreener, analysts’ consensus forecasts.
*Based on the 21 silicon chip firms in the SOX and excludes the nine semiconductor production equipment (SPE) companies that are in the index
**Assumes ARM generates flat sales and profits in the year to March 2024
*** Assumes $51 billion market cap for ARM
“The case for a premium valuation will rest upon five key points:
- ARM’s earnings are currently depressed by a slowdown in the smartphone market
- ARM’s dominant position in smartphone chip designs and its potential in other markets, such as AI
- The long-term growth prospects for the silicon chip industry, which has grown its revenues at a compound annual growth rate of 8% since 1980, and ARM’s own historic growth track record when it was listed in London between 1998 and 2016
- ARM’s fabless and chipless model, which means the firm should be highly cash generative and capable of generating very high returns on capital
- A net cash balance sheet
“A limited free float of some 10% could also make the shares illiquid and if the deal does catch fire, then the shares could be squeezed upward.
“The case against a premium valuation rests upon a different five points:
- ARM is locked in litigation with certain key customers
- Its sales are heavily dependent upon the maturing smartphone industry and a handful of clients generate a big chunk of sales
- China represents a quarter of sales and the combination of technological sanctions and Beijing’s economic woes is a concerning one (ARM China’s independent operations are also beyond management control)
- An offering syndicate of nearly 30 banks, which is so big as to look like an effort to quash any dissenting comment or research
- Softbank tried to set a valuation for the deal by buying in a 25% stake it did not own in ARM from the Vision Fund – a vehicle run by Softbank itself – that set an implied price of $64 billion, a price from which the market already seems to be shying away
“The motivation of Softbank is worth considering here. It is looking to sell the stake in ARM to reaffirm its status as a shrewd investor in technology firms after some embarrassing reverses, in the shape of WeWork and FTX, and a 30%-plus slide in its share price from 2021’s highs. It will be looking to prove to its own investor base that the $32 billion spent on the ARM takeover in 2016 was money well allocated.
“But in any IPO, there are three parties who are seeking satisfaction: the seller (who wishes to raise money and get value); the buyer (who wants to pay as little as possible and get value); and the syndicate running the deal (who wants to get paid and the kudos of a successful deal where the shares do well after the float and the after-market is well supported). Achieving the right balance between these three is not easy.
“Softbank looks to be going for a fat price, one that is hard to justify relative to the quoted comparables, or indeed ARM’s own history.
“At the peak of the technology bubble, ARM’s market cap peaked at £9 billion in 2000, when sales reached £100 million – or 90 times sales – a figure which swiftly proved unsustainable, as the tech bubble burst and ARM itself ultimately coughed up its first ever profit warning in 2002 as the smartphone market went through one of its periodic slowdowns. Investors are now being asked to pay broadly twice that multiple of sales, despite the company’s own history and the wider lessons of the 2000-03 tech bubble which were discussed by Scott McNealy, then the CEO of Sun Microsystems, a couple of years later, when he said:
“At ten times revenues, to give you a ten-year payback, I have to pay you 100% of revenues for ten straight years in dividends. That assumes I can get that by my shareholders. That assumes I have zero cost of goods sold, which is very hard for a computer company. That assumes zero expenses, which is really hard with 39,000 employees. That assumes I pay no taxes, which is very hard. And that assumes you pay no taxes on your dividends, which is kind of illegal. And that assumes with zero R&D for the next ten years, I can maintain the current revenue run rate. Now, having done that, would any of you like to buy my stock at $64? Do you realize how ridiculous those basic assumptions are? You do not need any transparency. You do not need any footnotes. What were you thinking?”
“Would-be investors in ARM are being asked to take an even longer-term view to get their payback, even if its business model looks capable of much greater longevity than that of Sun Microsystems, which was taken over by Larry Ellison’s Oracle for around $7 billion in 2010, way below its peak bubble market cap of $200 billion in 2000.”
