Exploring the IoT chip market, SoftBank's ARM acquisition
posted on
Aug 02, 2016 07:34PM
ARM Holdings plc has long been associated with the mobile boom; the chips designed and manufactured by the company are found in nearly every iPhone, iPad and Android device out there.
So why then would news that SoftBank intended to purchase ARM for $32 millionraise some eyebrows? SoftBank CEO Masayoshi Son justified the purchase, citing the growing potential of the internet of things and calling it a "paradigm shift" for the telecom and internet company based in Tokyo, Japan.
Yet the purchase raised another question: What exactly is ARM's position in the IoT chip market?
Cambridge, England-based ARM has had plenty of play in IoT; its processors are used in a number of multicontrollers in embedded systems; its Cortex-A series, for example, is used in smartphones, tablets and wearables from companies such as Apple, Samsung, Sony and Nokia. The company also has its own IoT device platform and IoT subsystem for its Cortex-M series, which is used in connected clothing, smart city, and virtual reality and augmented reality applications.
ARM has also made a number of IoT-related acquisitions over the past year, including hardware security IP and SoC software company Sansa Security, IoT security company Offspark, and imaging and embedded computer vision company Apical.
"ARM's position in the value chain extends its market longevity because it is already deployed 'embedded' in the products and business models of so many others vying for architectural dominance across connected devices," said independent analyst Jessica Groopman, who has covered IoT in various capacities for several years.
ARM's position will also help SoftBank, she said, giving the company a "significant head start" in markets such as wearables, drones, smart city applications and smart home devices.
Still, many have asked whether SoftBank's IoT expectations will be met, and whether ARM will see success in IoT as it has in mobile.
Mark Hung, a research vice president at Gartner, said the two cannot be compared; the IoT market differs greatly from the mobile market.
"Almost 2 billion handsets are shipped a year, and something on the order of 200 million tablets are shipped per year," Hung said. "For IoT to get to those numbers, it will take a while." He noted the oft-cited 21 billion device statistic, adding, "From the beginning of time to 2020, 21 billion devices will be deployed. Smartphones are by far bigger from a unit volume and revenue perspective."
However, Hung added that there is room for ARM to grow and potentially take over the space, though it is far too early to tell.
"I'll say ARM is in a good position," Hung said, "But I'm not saying that moving forward it's going to be prominent in IoT the way it is in mobile."
Understanding ARM's piece of the IoT chip puzzle -- and its future in it -- isn't always straightforward. After all, the company does not manufacture anything; it licenses its chipsets and, as an IP provider, receives royalties from companies producing the chips it designs.
"IoT devices today can have three basic types of processor architectures: ARM-based, MIPS-based or based on Intel's x86 architecture," Hung said. "There are a few based on very low-end proprietary microcontroller architectures, but those are probably very low-end and for higher-end IP processors. You can say today that the market is dominated by ARM, MIPS andIntel. And of those three, ARM is definitely the leader, at least on the IoT end."
However, unlike in mobile, IoT has a number of varying use cases, which may not bode well for ARM or any other vendor wishing to lead in the space.
"A smart sensor that goes out in the field for energy exploration purposes is going to be very different from what's required in your smart refrigerator," Hung said. "What an IoT device needs to do for manufacturing is very different from one needed for retail. Because of the diverse set of applications for IoT, I think it's going to be hard for ARM to dominate ... the plethora of architectures [for IoT] is going to dwarf that of mobile devices."
You can say today that the market is dominated by ARM, MIPS and Intel. And of those three, ARM is definitely the leader, at least on the IoT end.Mark Hungresearch vice president, Gartner
That's not to say semiconductor companies won't see success in the IoT chip market, Hung said, offering the example of autonomous vehicles and explaining that though there are a fewer number of cars sold each year, the chips that go into them are more expensive.
"I would say that segments of the IoT market will definitely attract vendors to start," Hung said. "Then it's going to over time start permeating across other industries as well."
"From an IP perspective, ARM is the dominant player," Hung said. "But if you look at the semiconductor vendors themselves, Intel has invested a lot in the space, and other chipset semiconductor giants have invested a lot as well."
Sharon Shea asks:
0 Responses
Join the DiscussionHung mentioned companies including Qualcomm, Texas Instruments, STMicroelectronics and NXP.
"Up and down the food chain, whether it's by size or marketplace, semiconductor vendors are definitely investing a lot in this space," Hung said. "A lot of them do use the ARM architecture, so in a sense it's good for ARM. But Intel is definitely to be reckoned with in the future as well."
Groopman noted that many of the M&As over the past few years have been related to IoT, especially when it comes to the semiconductor space.
The current state is shifting ownerships, alliances, business models ... but none of this spells maturity yet.Jessica Groopmanindependent analyst
"We've seen a huge consolidation in this market with mergers involving SanDisk acquired by Western Digital; KLA-Tencor acquired by Lam Research; CSR acquired by Qualcomm; Altera acquired by Intel; Freescale Semiconductor acquired by NXP. The list goes on and on," Groopman said. "The current state is shifting ownerships, alliances, business models ... but none of this spells maturity yet. These constituencies are all vying for the central role in powering IoT devices."
The ARM acquisition news isn't a bad thing, Hung said, and may even be beneficial for both ARM and other vendors.
"I think one good thing about this purchase is because SoftBank did not really participate in the semiconductor market before and it's pretty much leaving ARM management in place and letting it continue with its business model, that speaks well to ARM's future prospects, in that its current customers will not feel threatened by the acquisition," Hung said.
"Put another way," Hung added, "If someone like Apple or Samsung or Qualcomm or Intel was to buy ARM, I think ARM would lose a lot of business because existing competitors to those companies would definitely change their product rollouts to not use ARM in the future. But because SoftBank can be properly viewed as aneutral third party, I think the acquisition should not impact ARM's future prospects negatively and, in fact, may perhaps provide them with more financial resources going forward to compete with the other companies in this space."