Updated: 25th February 2018

Intel Kills the Atom SoC Powering Microsoft’s HoloLens


Last month, we wrote about fresh information regarding the followup to Microsoft’s HoloLens, and also how the new device would integrate an AI co-processor. At the moment, we speculated that Intel’s Atom SoCs may not be very long for this world, although we didn’t anticipate the hammer to fall quite as quickly as it has. Intel is murdering its own Atom X5-Z8100P processor, the chip that powers the Microsoft HoloLens.

This chip is not publicly recorded on Intel’s website (Ark.Intel.com; helpful for several CPU identification jobs), but was identified (PDF, but with security caution)  as an Cherry Trail-based component clocked in 1.04GHz. Intel’s decision to cancel its Atom-based SoCs since it exited the smartphone and tablet computer business has left Microsoft without a followup solution, although the fact that Intel is leaving the market for a chip which may only have had one client suggests Microsoft and Intel have worked this issue out amongst themselves.

Another motive to view the situation as a mutual arrangement is Intel hasn’t murdered its other Cherry Trail Atom SoCs at the same household:

Lots of Intel Cherry Route SoCs are still on the market.

It raises the issue, however: If Intel is not providing a CPU solution for HoloLens two (our title), who is? We might not understand which firm, specifically, is building the alternative, but there are only 3 options, and only one of these makes sense:

1). Microsoft has paid Intel to build it a semicustom Atom SoC according to Goldmont.
2). Microsoft compensated AMD to build it a semicustom SoC according to a 14nm Jaguar core.
3). Microsoft contracted by an ARM vendor.

#3 is the only sensible alternative here. Microsoft might have compensated Intel for a habit Atom SoC earlier, but this was back when Intel was still committed to selling Atom processors across the tablet computer and smartphone markets, including in Microsoft’s own Surface 3. It’s one thing to become a semicustom client for the mildly custom part that’s being sold everywhere, and also another thing entirely to be the only client the company expects to get for an entire hardware line. MS can scarcely hope to create its investment back from bulk; there is no sign that the organization is planning a broad customer rollout for two.

The second choice does not make sense either, for the same reason. We know AMD has ported Jaguar to 14nm, but the SoC designs for the PlayStation Professional and Xbox One X do not possess a mobile-compatible GPU bolted on. And AMD has created only the faintest of overtures towards the pill market because it offered the GPU IP which became Qualcomm’s Adreno almost ten years ago. Again, Microsoft is not likely to want to foot the entire bill for a cheap x86 SoC and associated graphics hardware, especially when AMD, like Intel, has no plans to compete at the tablet or smartphone space.

That leaves choice #3. Here, there are an array of options and a lot of companies who would love a high-profile win at a device such as HoloLens. Both Qualcomm and Imagination Technologies possess GPUs which will deal with the processing, especially considering that Microsoft is beefing up the function and capability of its own artificial intelligence processing.

This last point is presumably why the company wouldn’t turn to a company like Nvidia. While Nvidia could involve some silicon in its own arsenal that could be repurposed for this function, strapping Microsoft’s habit AI silicon into precisely the identical hardware which features Nvidia’s AI ion (via the GPU) will be redundant. Besides, Nvidia has gotten from tablet computers and smartphones as well, and is focused on vehicles, which have a great deal more robust cooling choices and space for larger boards than your average headset.

Having a Microsoft announcement that it is getting into the CPU business, is the only probably match.

This website may make affiliate commissions from your hyperlinks on this page. Conditions of use.