Arm – Table debate

SoftBank takes most of the value.

• At $51 per share, investors are being asked to pay for much of the growth that the company says it should be able to generate meaning that there is not very much being left on the table for public market investors.
• Despite the current malaise in the smartphone market, Arm is expecting a fairly rapid resumption of road growth combined with margin expansion until at least 2026 (see here).
• At a high level this looks to be pie in the sky given the current environment, but breaking down the components reveals how Arm might get there.
  • First, revenue mix. Most observers link Arm directly with the smartphone market, but according to the F-1, less than half of Arm’s revenues come from the smartphone market.
  • This is due to the weakness of that market and the strength of other markets like automotive into which Arm has successfully diversified.
  • This has reduced the negative impact of smartphones on Arm.
  • This can be seen in the financials where FY 2023 and Q1 2024 revenues fell by far less than the smartphone market.
  • Second, price where new processor cores with more technology built into them has allowed Arm to increase the price that it charges for its cores.
  • Arm’s biggest customers like Apple, Nvidia and soon Qualcomm design their own processors and so will pay lower royalties, but a large part of the market will.
  • As more advanced processors trickle down into the mass market from the premium tiers, this should result in a steady increase in price for Arm over a multi-year period.
  • This has gone down with some grumbling from the industry but compared to building it all oneself, the argument that Arm is providing value is fairly strong.
  • Third, business model change where Arm could change its model from charging semiconductor companies to charging royalties on the end device.
  • In its responses to Arm’s lawsuit, Qualcomm has alleged that Arm intends to change its business model from charging royalties on chipsets to charging royalties on end devices.
  • This is the same business model that Qualcomm has pursued with great success over the last 30 years.
  • On a smartphone market worth roughly $500bn, Arm currently earns around 0.2% which would increase greatly if its business model was to change.
  • If Arm could charge 1% of the device’s price rather than a percentage of the price of the chip, then its revenues would increase by 5x even if the market were to remain flat.
  • Arm has emphatically denied that any change in business model is being planned, but does not completely rule it out in the risk warnings in the F-1 (page 37) although this refers to negative changes rather than positive changes.
• Putting all of this together and looking at 2026, a blue-sky revenue forecast could be $6.7bn ($5bn mobile phones (business model change) & $1.7bn other growing in line with expectations).
• With 60% operating margins and 20% tax, this would give net income of around $3.2bn which would translate into EPS of $3.14 or 16.2x 2026 PER
• If growth is going to be steady, then a PER multiple of 25x – 30x could be applicable giving a valuation of $78 – $94 or an annualised return of 25%+ from the IPO price of $51.
• This is a pretty good deal, if one believes the blue-sky scenario.
• The biggest question mark here is business model change as it will be required to be successfully executed to achieve this blue-sky scenario.
• This combined with the rising risks in China where Arm earns a lot of revenue but has no control over those revenues and the competitive threat of RISC-V eroding the low end asks a lot of public market investors.
• Consequently, I am not convinced that there is very much being left on the table for investors at an IPO price of $51 leaving us with the AI story.
• Closing the books early and pricing high indicates to me that the public market investors are thinking of Arm as an AI company like Nvidia making the high multiples much more palatable.
• The problem is that there is no evidence that I have seen that leads me to believe that the current hype on AI is not overblown.
• There are significant use cases for generative AI, but I have not seen anything that leads me to think that we are at the beginning of a superintelligent AI era.
• Hence, when expectations fail to be met and start-ups come back to the market for more money, disappointment will set in, valuations will fall, and the market will move on to the next new and shiny theme.
• The problem that Nvidia has is that if the bubble pops, its valuation is likely to get crushed along with any other company that has been sold on the back of this theme.
• This is why despite the blue sky scenario, I am not convinced that SoftBank is leaving much on the table on a risk-adjusted basis for those buying in now.