ARM is now the dominant microprocessor architecture, a feat achieved through a combination of sharp engineering, a PC market failure, and an inspired business model pivot. The reward is a puzzlingly high acquisition offer from Masa Son, the maverick SoftBank founder.
Chips built on the Advanced RISC Machine architecture are everywhere. Available in a wide range of sizes and power requirements, ARM processors impart intelligence and connectivity to devices ranging from high-end smartphones, tablets, watches to thermostats, drones, infotainment systems, and many other appliances. According to Wikipedia’s ARM Holdings article, 15 billion chips containing an ARM core were sold in 2015.
How ARM managed to dethrone the previous emperor, Intel, makes for a fascinating story that can be summarized in three idiosyncratic, illuminating steps that include one ingenious move:
- Founded in the UK in 1978, Acorn Computers started life as a PC hardware company. Initially successful, Acorn couldn’t compete with IBM in the business sector.
- The company extracted a choice bone from the corpse: The Acorn RISC Machine chip.
- In a splendid pivot move, Acorn dematerialized the bone: They turned ARM into a licensable design.
Along the way, two heroes stood out:
- Sophie Wilson designed the instruction set for the first Acorn RISC Machine and received a lifetime award for her work.
- Hermann Hauser was instrumental in spinning off ARM Holdings from Acorn.
(The links above will give you many more details on the complicated life of Acorn Computers, its struggles, and its innovations.)
But the ARM story needed another ingredient—an enabling innovation without which the heart of ARM’s business model, the licensing of processor design, would be impossible: Electronic Design Automation (EDA).
As integrated circuits grew to comprise thousands and then millions of logic elements, breadboards were virtualized: The circuit-to-be was designed on a computer, just as we model a building using architectural Computer Assisted Design (CAD).
A multibillion dollar industry of software modules that could be plugged into one’s own circuit specifications soon emerged. Companies such as Synopsys, Cadence, and Mentor Graphics offered circuit design tools, and an ecosystem of third-party developers offered complementary libraries for graphics, networking, and sensors. The end result is a System On a Chip (SOC) that’s sent off to semiconductor manufacturing companies commonly called foundries.
This was the fertile ground on which ARM has prospered. ARM-based chips aren’t simply more efficient and cheaper than Intel’s x86 designs, they’re customizable: They can be tuned to fit the client’s project.
Intel didn’t get it. “Just you wait!” the company insisted, “Our superior semiconductor manufacturing process will negate ARM’s thriftier power consumption and production costs!”
But that opportunity has passed. Intel miscalculated the iPhone, failed to gain any traction in the Android market, and had to resort to bribing (er… “incentivizing”) tablet manufacturers to use their low-end Atom processors. Earlier this year, they threw in the towel on mobile and are now focused on PCs and Cloud data centers.
The flexibility, cost, and power budgets of customizable ARM processors won the day, leading to 15 billion ARM-based chips in 2015.
Today, ARM Holdings is a $1.5 billion company with +15% year-to-year growth, nice financials (such as 96.7% gross margin), and a 46.7% operating margin. (For all the details—perhaps too many—see this 2015 presentation.)
15 billion ARM-based chips for $1.5 billion revenue means that, on average, ARM gets a licensing revenue of 10 cents per chip, and spends a little less than of half of that, 4.7 cents, to generate such revenue. It sure beats today’s Windows PC business and its measly 5% to 7% operating margins in the best of cases.
ARM Holdings is doing well by any measure, but the price paid for an asset is supposed to reflect expectations of future gains. If we compound ARM’s 15% revenue growth over ten years we come up with $6 billion, a four-fold revenue increase… so why is SoftBank willing to pay $32 billion, more than 20 times current revenue?
Masayoshi (Masa) Son, SoftBank’s founder, isn’t a wide-eyed, newly-rich entrepreneur looking to make—or lose—a quick buck. Having built a large PC software business, he diversified into telecoms in Japan and the US (SoftBank owns 80% of Sprint). He also holds a piece of Yahoo! which, if this weekend’s rumors are true, is soon to be rolled into Verizon.
Having survived the 2000 dot-com bubble and the 2008 financial crisis with a personal fortune estimated at $17 billion, Son isn’t shy about criticizing the short-term views of US investors and their fixation on “shareholder value” at the expense of other longer-term metrics and societal contributions.
With his accomplishments and maverick attitudes in mind, we must conclude that Son sees many more ARM-based chips in our future, some with a revenue-per-unit that’s higher than today’s 10 cents.
But where does he see them?
The Internet of Things era is emerging more slowly than its enthusiastic evangelists had predicted, but it is happening. Look at all the objects around us that sooner or later will sport ARM-based intelligence; this will result in huge numbers, tens of billions of chips. Of course, these will be low-end chips due to power and cost restrictions, which will likely lower the average per-chip revenue, not increase it (let’s not forget that ARM doesn’t sell chips—it makes its money on licensing).
The Smartphone’s go-go years are behind us. Smartphones will continue to be a healthy business, but there will be some downward pressure from manufacturers in a race to the bottom of the price range. The same applies to “pure” tablets.
Some hope could come from tablet/PC “toaster-fridges.” As high-end ARM chips grow to provide increased computing power, the hybrid combos could move away from Intel designs. Or perhaps Apple will cannibalize its Macintosh line with ARM-based 2-in-1 devices. Perhaps. In any event, it’s a replacement game in a PC market that’s long been in decline.
The last fortress is the Cloud and its data centers—Intel’s best money-making segment. It could be the perfect fit: Cloud machines mostly run ARM-friendly Linux derivatives, and as data centers become more energy conscious and ARM chips more powerful, an opportunity for more unit volume and higher prices could present itself. Still, Cloud server volumes are in the tens of millions per year, not billions as in smartphones.
None of these opportunities, from ubiquitous but low-end IoT devices, to high-end but low-volume Cloud servers, promise higher revenues and profits… which might explain the generally expressed skepticism regarding the high $32 billion offered for the ARM Holdings acquisition.
Or maybe Masa Son is betting on a completely different factor. ARM has a de facto monopoly position on SOC and embedded applications. This leads to pricing power, an aspect that is, according to Warren Buffet, “the single most important decision in evaluating a business.”
Simple in theory, just turn the screw. Harder in practice when the largest volume opportunity is in the lowest price segment.
The ARM story will continue to be interesting.
This post originally appeared at Monday Note.