Intel’s Battle with Arm

Unless you have followed Intel for at least 10 years, you might not now know that at one point, Intel actually owned an Arm business. Up until 2006 Intel owned XScale, an ARM-based chipmaker that it sold that year for $600 million.

Intel sold XScale because it wanted to double down on the x86 architecture that had made it so successful. Intel was working on a low-power version of x86 chips called Atom., Their leadership at that time believed that selling ARM chips would show a lack of commitment to the Atom platform.

But this decision, that looks like a really bad one now, can be traced back to the early 2000’s when then Intel CEO, Craig Barrett, had a series of meetings with Intel’s leadership in Israel, who convinced him they could create a low voltage X86 chip that could compete with Arm. At the time, X86 processors ruled the PC world and Intel was convinced that Windows would continue to dominate in PCs as well as any type of mobile platform that could arise in the future.

I was at an event at Intel just after Barrett had committed to the proposal from Intel’s Israel tech center and asked him directly about his thoughts on Intel’s ability to create a low voltage processor. He told me that Windows would be the dominant OS for decades to come and that he was assured by his team that they could create a processor that could meet and eventually surpass ARM in the low voltage race.

At the time that it seemed like a safe bet and indeed, Intel accelerated the work on their X86 Atom chips, although the time frame Intel Israel suggested they could deliver a really low voltage processor was off by at least 7-10 years at best.

In the meantime, various things have happened that made this bet by Barrett and subsequent Intel CEO’s miss the smartphone revolution and could, in the near future, impact even their traditional PC business.

The first thing that happened was that Intel could not have guessed that Apple would create their own mobile OS, or that Google would jump in with Android and together dominate the world of smartphones operating on an ARM core. Even worse for them, when Jobs asked Intel CEO Paul Otellini in 2006 to create a chip for the iPhone, the standard line has been that Otellini did not see the future of smartphones and declined. But in truth, the real reason had more to do with the Atom being too power hungry and could not deliver on the specific requirements that the iPhone team needed to run a very lightweight OS and could last at least 24 hours using a pretty small lithium-ion battery.

The second thing that has happened is that Microsoft has evolved their OS so that it no longer has to run just on an X86 processor. Microsoft too lost the smartphone wars by doubling down on Windows on X86 and they did not understand the role ARM processors would play in the evolution of mobile computing. But that is changing. After a false start with their Windows on Arm program a few years back, Microsoft and Qualcomm have renewed their partnership and is moving rapidly to make Windows OS work much better on Arm.

While the X86 version of Windows is still the best in terms of delivering performance against a Qualcomm 835, 845 or 850 processor, various reports suggest that Qualcomm has a new chip in the works, according to Ars Technica, that could give Intel a run for their money-

“The Snapdragon 1000 is believed to be an even more powerful laptop chip intended to go head to head with Intel’s Y- and U-series Core processors. These have a 4.5W and 15W power envelope, respectively, and are used in a wide range of tablets and Ultrabook-type laptops. The Snapdragon 1000 is reported to have a 6.5W power draw for the CPU itself, with a total power draw of 12W for the entire SoC. The Snapdragon 1000 test platform has 16GB of LPDDR4X RAM and two 128GB UFS flash drives. It also has 802.11ad gigabit Wi-Fi, gigabit LTE, and a new power management controller.”

While we have not been briefed by Intel on any new low voltage processors in about a year, you can bet that they are not sitting on their hands and are working hard to get too much lower voltage that can deliver longer battery life, which appears to be the main objective of this new processor from Qualcomm.

On the other hand, the new generation of Arm processors are being manufactured in 7 nm while Intel’s are still being made using 10 nm and 14 nm processes and Intel is possibly at least two years out from using 7nm processes in their own fabs to be able to get even greater performance and battery life on their newest mobile chips.

To see if this threat is real to Intel I have been testing one of Lenovo’s new Qualcomm 835 based ACAO (Always Connected-Always On) 2-in-1’s for the last two weeks. I even kept using Windows 10S instead of upgrading to the full Windows version and have been surprised at acceptable performance, great battery life and the fact that it worked relatively smoothly with all of the apps and web apps I use daily. I did have some performance glitches but nothing that would keep me from being productive.

Since we have not had any updates from Intel for some time about processor roadmaps, we can only speculate that they are working to counter any new Qualcomm push into PCs. We also do not know any real specifics on the reported 1000 so at this point, it is hard to assess how competitive it will be.

But, I often wonder where Intel could be today had they not sold off XScale and could have filled Steve Jobs request to deliver a chip for the original iPhone. Or, at least, been able to sell chips to Samsung and other phone makers who today use chips exclusively based on Arm processors.

Although there are a lot of unknown’s still about the upcoming battle between Intel and Qualcomm in PCs, I have no doubt that Qualcomm, with Microsoft support, is committed to this but that Intel, with cash and major incentives, will try and thwart any advances from Qualcomm to keep their dominant place in PCs.