Arm plans to spin out its IoT Platform business and its 2018 Treasure Data acquisition to new entities owned by parent company SoftBank so it can focus on its core chip design business. The chip design company, whose IP is in almost all of the world’s smartphones (and soon in Apple’s MacBooks) said on Tuesday that the proposed transactions would let SoftBank, which has more experience managing growing businesses, take on the fast-growing IoT services provided by those two business groups. I think that’s the nicest way of saying that Arm has failed to deliver what the IoT market wanted even though it was one of the first companies to see where market needs and demand were headed.
In a press release from Arm, CEO Simon Segars says, “Arm believes there are great opportunities in the symbiotic growth of data and compute. SoftBank’s experience in managing fast-growing, early-stage businesses would enable ISG [IoT Services Group] to maximize its value in capturing the data opportunity.”
For the last eight years, Arm has seen the internet of things coming, and prepared itself through a variety of acquisitions that range from power-saving chip technology to low-power machine-to-machine communication protocols. It even proposed its own cloud platform for IoT back in 2016. And in 2018 it purchased Treasure Data which offered a full management platform for IoT and launched Pelion, a full-service IoT management platform for companies that want to build connected products and services. I’ve actually spoken to several happy Pelion customers, but their happiness had nothing to do with ARM’s core chip business. And when it came to ARM’s own Mbed cloud efforts, ARM didn’t move the needle.
The original hope was that ARM could link its devices and cloud services in ways that we’re finally seeing companies such as Amazon and Microsoft do with chip vendors. Creating established libraries, software development kits, and even wrapping licensing agreements for a chosen OS to particular chip sale are making it easier for developers to secure, build on and send data to their respective clouds. Even smaller providers such as Ayla Networks and Tuya have signed deals with chip vendors to make it super easy to connect to their platform and pull data from a device. Those links also help with security as one can see from Microsoft’s focus on Azure Sphere.
Many of these deals involve chips that use Arm-based microcontroller designs and even Arm-based application processor designs, which is a possible reason that Arm struggled to sell its own cloud + hardware vision.
First, why use an Arm cloud, when your data and developer expertise is built on another cloud platform? Second, why would Arm push too hard to tie its cloud to its designs and risk alienating its actual end customers who want to use their own cloud tie-ins as a point of potential competitive differentiation?
Arm knew what the IoT needed, and it was working hard to pull together all of the pieces, but it was up against both established competitors and a classic innovator’s dilemma. This is a challenge for a lot of the semiconductor companies that want to move up the stack from providing commodity hardware (that requires a large amount of expensive software development expertise that they can’t charge for) to offering services and software that could generate additional revenue. Arm doesn’t make the chips, but like a chip company, it has to sell further up the stack and risk alienating its customers if it pushes too hard on software or services those customers want to add.
So spinning out these offerings to an independent business makes sense. Plus, SoftBank is looking to exit some of its IoT investments, and if it can create a compelling platform that actually has existing customers, it could provide an exit for SoftBank in the not-too-distant-future.
As far as the big picture for Arm, this deal isn’t Arm giving up on the IoT; rather it’s Arm giving up on the services side of IoT. Arm will still design processors that sit inside phones, routers, smart speakers, and more. Even so, I am on the lookout for RISC-V architectures and other new chip designs that are gunning for market share in the IoT and could tear business away from Arm.
In today’s digital world, no one can afford to sit still; and even when a company can see what’s coming, it doesn’t mean it gets to deliver it.
Steve Lurie says
Kind of also raises the importance of standards, like LwM2M. If you build on standards, you can easily move away from Pelion, as some clients will do. Otherwise, you are stuck with whoever Softbank eventually sells Pelion to, or worse yet, what happens if they drop it? Check your LwM2M compliance using https://iowa-server.ioterop.com/device