SoftBank is seeking an equity value of $50 billion to $54
billion as part of the roadshow for Arm’s initial public offering, Reuters
reported at the weekend. When the Japanese conglomerate’s boss Masayoshi Son
scooped up the Cambridge-based company seven years ago, he was placing a bet on
the so-called Internet of Things, the view that tiny semiconductors would
proliferate across everyday household items and provide the next leg of growth
for chip designers. Now he’s pitching it instead as a central player in
artificial intelligence, the new industry vogue.
Stock-market AI darling Nvidia (NVDA.O) at first glance
seems to validate Son’s valuation hopes. The Californian chip designer’s
enterprise value has recently soared to $1.2 trillion, or 23 times analysts’
average forecasts for its revenue this year, LSEG data shows. On that basis,
Arm’s top line could stagnate at $2.7 billion and still justify a greater than
$60 billion equity value, after including its $2 billion of net cash.
But that misunderstands Nvidia’s ascent, which is
underpinned by massive increases in forecasts for operating profit. The mean
analyst prediction for earnings before interest and taxation in the financial
year to January 2026 has more than doubled since the start of May. The lesson
for SoftBank and Arm is clear: chip investors are laser-focused on medium-term
operating profit, not just revenue.
There are three key moving parts to Arm’s valuation. First is its top-line growth, which has been lumpy, dropping 1% in the last fiscal year but leaping 33% in the preceding 12 months. SoftBank in a recent earnings presentation highlighted a longer-term trend: Arm’s 14% compound annual growth rate over the past three years, using fiscal first-quarter figures. Apply that pace of expansion to the most recent calendar year, which makes it easier to compare Arm to listed peers, and the company’s revenue would reach $4 billion in 2025.
Second is the operating margin, or the percentage of sales
that’s left over after deducting all costs other than interest and taxes. Under
SoftBank, Arm’s operating margin has dropped to around 25%, from roughly 40% in
2015 – a consequence of Son’s preference for heavy investments in research.
Bernstein analysts reckon a publicly listed Arm might nudge the margin back up
to 33% within three years, if it lets costs grow slightly slower than revenue.
On that basis, the company would pump out $1.3 billion of operating profit in
2025.
Third is the valuation multiple, which is best judged by
looking at what investors are prepared to pay for similar listed companies.
Nvidia, Cadence Design Systems (CDNS.O) and Synopsys (SNPS.O) are good choices.
The first is exposed to Son’s beloved AI trend, while the latter two U.S. firms
make software for chip designers, much like Arm, which licenses intellectual
property for processors. On average, the trio have a current enterprise value of
25 times the operating profit analysts reckon they’ll earn in 2025, using LSEG
data.
If Arm nabbed the same multiple, its enterprise value would
be $33 billion, using the above growth and operating margin. Add the SoftBank-owned
company’s net cash as of June 30, and the market capitalisation would be $35
billion.
Son last month agreed to pay a significantly higher
valuation of $64 billion when SoftBank bought 25% of Arm from the Saudi
Arabia-backed Vision Fund. In that context, the new $50 billion to $54 billion
target would already represent an embarrassing step down. But even to justify
the lower valuation, investors would have to torture the assumptions.
Start with revenue. AI star Nvidia’s top line will grow at a
compound annual rate of 51% from 2022 to 2025, according to average analyst
forecasts gathered by LSEG. Assume that Arm captures some of the same gold dust
and grows at a 20% annual clip up to 2025. Holding other assumptions constant,
Arm’s fair value would rise to $41 billion. It’s hard to see a case for raising
the valuation multiple beyond 25 times, which is already much higher than the
average 22 times multiple that Arm managed in its final three years as a public
company before SoftBank swooped. So profitability has to do the final lift. To
clear a $50 billion equity value the 2025 operating margin would have to
surpass pre-SoftBank levels of around 40%, while revenue grows at 20% a year.
The company itself acknowledged this tradeoff in a 2021
filing making the case for its sale to Nvidia, which was blocked by antitrust
authorities. “The capital markets would expect Arm to make significant
strategic changes, including cutting costs…” the parties said, referring to the
possibility of an IPO, concluding that the UK group faced “significant
challenges to growth” as a standalone company. SoftBank’s hoped-for price range
only works if investors are willing to believe two mutually contradictory
things at once. That may be possible for a short while, especially if the stock
market’s AI hype persists, but it’s hardly the basis for a stable valuation
over time.
SoftBank Group’s chip designer Arm is planning to ask
investors to pay between $47 and $51 for each of its shares when it begins
marketing its initial public offering, Reuters reported on Sept. 2 citing
people familiar with the matter. The price range would translate into a
valuation for Arm of between $50 billion and $54 billion, according to the
report.
Customers of Arm including Apple, Nvidia, Alphabet and
Advanced Micro Devices have agreed to invest in the chip designer's initial
public offering at a valuation between $50 billion and $55 billion, Reuters
reported on Sept. 1 citing people familiar with the matter.
SoftBank agreed in August to acquire the 25% of Arm that it
did not already own from its Vision Fund investment vehicle at a valuation of
$64 billion.
Arm plans to price its shares on Sept. 13, with trading
scheduled to begin on the Nasdaq exchange the following day, according to
Reuters.