Why SoftBank wants ARM (And ARM needs SoftBank)

ARM’s CEO Simon Segars explains the deal here.

ARM Holdings PLC, the British chip design firm with technology that’s inside almost every mobile phone, has accepted a £24.3 billion ($32 billion) buyout by Japan’s SoftBank. It awaits various approvals.

The deal came as a shock on Monday morning. It was announced a few weeks after the UK had voted to leave the European Union, while the British pound was in freefall. It also came a month after the departure of SoftBank’s CEO heir apparent Nikesh Arora.

I can’t speak to the political aspects of this deal. Only about technology. And here it’s clear that SoftBank is trying to cash in on the future of technology. As SoftBank CEO Masayoshi Son said in a statement associated with the deal:

“ARM will be an excellent strategic fit within the SoftBank group as we invest to capture the very significant opportunities provided by the Internet of things. This is one of the most important acquisitions we have ever made.”


This deal is good for ARM, which is getting an excellent price for its shares and the potential of billions more in capital from SoftBank. But I’m not sure SoftBank has much of a huge strategic reason.

Son is a huge believer that the combination of computing everywhere and artificial intelligence will bring about the singularity when artificial intelligence becomes smarter than humanity. In an interview explaining why he decided to stay on as CEO of SoftBank, Son told the Nikkei Asian Review:

“I think we are about to see the biggest paradigm shift in human history. The Singularity is coming. Artificial intelligence will overtake human beings not just in terms of knowledge, but in terms of intelligence. That will happen this century.”

For those less enamored of this besting of man by machine, forget the science fiction aspects and look around you. People turn to computers to augment their decision-making all the time. Already, people check Yelp on their phones before deciding on a restaurant. You probably check Google or ask Siri a question dozens of times a day. And as the ability to gather more data cheaply and mine it for insights improves and costs less, you’ll see this play out in business (see my story last week on asking Amazon’s Echo for sales data) and in your personal life.

Mark Rolston, the CEO of product design firm argodesign, has told me that he believes that in the near future people won’t make any decisions without a dose of AI. In many cases the phone will be the delivery device for that infusion of data, but it could very well be machines acting on our behalf. One example might be a thermostat tied to your bank account trying to keep to a set budget. Another would be a nutrition system that connects to your activity tracker and your pantry to suggest the optimal meal.

The key for all of these devices (and the phone) is the chip inside. It’s an ARM-based processor inside the thermostat, the phone and the activity trackers.

ARM was created from a collaboration between Apple and Acorn, which built a special processor for the Apple Newton PDA in the early 90s. As the mobile industry grew, ARM licensed its chip designs to a variety of chipmakers such as Qualcomm and Samsung.

But ARM isn’t just about mobile, which is where many of the commenters on today’s news are focused. (This makes sense since SoftBank owns a Japanese operator and Sprint.)

As computing grew more energy intensive ARM also saw a chance to break into the world of servers, challenging Intel. Its licensees grew to include companies like Applied Micro and AMD, with many of them building specialized silicon on top of the ARM design. ARM’s server fight is just beginning, but Google, Amazon and others have an interest in using ARM-based chips.

ARM also has several designs in the networking and cellular equipment arena with its Cortex A57 and Cortex A72 designs. Intel has also branched into this space of providing chips for cellular equipment makers. Essentially, ARM is trying to move upmarket into servers and more processor-intensive cellular base station gear, maintain its existing middle market in the cell phone world, and focus intently on the low-end market for microprocessors that are used in smart watches, appliances and other devices that have historically not needed a lot of intelligence.

ARM has to do this. While the low-end microprocessor market doesn’t have high average selling prices for the chips, the industry expects to sell a lot of them as more devices become computers. On the mobile phone side, the market for handsets is saturated, but ARM still needs to invest in new designs that can handle upcoming trends like artificial intelligence. AI capabilities will also play a huge role in the server market.

In short, this is a bet on infrastructure. Not just in the obvious mobile and embedded computing space that ARM has known and mastered. If SoftBank’s plans for advancements of ARM’s technology are to be believed, ARM will will have the capital heft to invest more into server technology and improving artificial intelligence capabilities on all of its chips. It can also spend on security in the server and cellular equipment market, which will give it the so-called “end-to-end” hardware-based security that Intel likes to discuss.

For the last eight years, I’ve been keen to see how ARM handles its growing success and its challenge to Intel. Intel designs and makes its own chips using what’s called the x86 architecture (Intel licenses that architecture to AMD) while ARM merely designs an instruction set it licenses to a variety of chipmakers (andincreasingly others). The implications of this are significant.

It means a company building on ARM can differentiate its silicon in a way that Intel cannot. As software eats the world, this layer of flexibility for the buyers or users of hardware gives ARM a subtle edge that may be as relevant as its power efficiency.

This also made ARM a tough target for a buyout. Because so many companies in the chip world are customers of ARM, it couldn’t sell to another chip company without risking the loss of its business. ARM has always been about staying a neutral ecosystem provider. Even Google might have a hard time making an ARM buy since other data center customers investigating ARM-based silicon might look askance.

SoftBank is a different story. The conglomerate owns SoftBank and Sprint, which are two mobile carriers, but it also has investments in e-commerce, gaming companies and even owns a publisher. What’s unclear is if SoftBank wants to be a neutral ecosystem player. Essentially an arms dealer for computing. If so, it could make the investments needed to keep ARM competitive as what Son thinks of as the Singularity approaches.

I’ll delve more into what those investments should be in our regular newsletter after I’ve had more time to think about the implications. In the meantime, this deal is a shock, only because it was hard to see anyone buying ARM. It’s also unclear how well-equipped SoftBank is to take advantage of this acquisition.

However, if SoftBank is willing to invest beyond the buyout price, it could have a perfect asset for the future of computing.

This was a special edition of the Stacey Knows Things newsletter.

Did you like this story? Want more? Sign up for Stacey Knows Things, a newsletter covering the internet of things, to get this essay and more.

I blog about chips, broadband and the internet of all the things.