Intel Throws in the Towel on the Atom Processor

[dgstorm]

Sorry for not posting this here sooner. We had it up on our sister-sites, but accidentally missed posting it here. Even though this news is a couple weeks old, it's worth sharing, just in case you missed it previously.

This was a bit of a shock from out of left field late yesterday. Intel announced they will be cancelling their SoFIA platform and the next gen Broxton chips. That basically means they have effectively killed the Atom processor, which has been struggling to make any headway in the mobile world.

Intel did not clarify if this means they are exiting the mobile market altogether or if they are simply planning to move in a different direction with some unknown future tech. The principal reason for Intel's lack of success in the mobile chipset market isn't even just the heavy competition from Qualcomm, Samsung and MediaTek in the space. The fact of the matter is that the Atom chipset never lived up to Intel's media hype in performance or battery efficiency.

Now we are left wondering about future and current tech that utilizes the Atom processor. What does this mean for various mid-range smartphones and the upcoming Microsoft HoloLens? Apparently, they will need something new to power them in the future. Will it be Samsung, Qualcomm or someone else entirely?

For Microsoft at least, retooling the HoloLens for something shouldn't be too tough. In fact, most of the current rumint indicates that Microsoft was already planning on using Qualcomm's upcoming Snapdragon 830 for the Surface Phone. If they are already partnered up with Qualcomm on this device, it seems like it would be easy to replace the internals of the HoloLens with a Qualcomm chipset. This is especially true considering it is still in the prototype stages.

What do you folks think? Is Intel getting out of the mobile market, or will they come up with something to replace the Atom?

Source: SlashGear

[zeo]

Well, we can point to a few articles from Anandtech that shows they did pretty well on the power efficiency, it was mainly the performance that they fell short on but the real issue was they couldn't match costs... While the ATOM was pretty cheap for them to make it was still not as cheap as ARM SoCs and lacked the ability to easily customize and optimize for specific OEM needs.

Intel tried to compensate with heavy subsidizing but it was costing them billions and they only garnered about 1% gain of the mobile market for their efforts and as they started to phase out subsidizing partners like Asus started to move away from Intel based mobile devices and back to ARM based products.

Add, the tablet market has started to decline the last 2 years left Intel with a ever shrinking prospect in the mobile market that wasn't going to change any time soon... So it appears they're cutting their losses for now.

So this is what's known, Intel is officially dropping out of the Smart Phone market but existing products will continue to be produced, they just won't be updating them... Since Broxton was going to server not only tablets but also Smart Phones, it is what has been cancelled, along with the phone SoC specific SoFIA updates.

The ATOM architecture is not dead, however, as the 2 in 1 market and low cost laptop market is still thriving and thus the other half of the Goldmont update is still being released with Apollo Lake...

Now, this is essentially the same architecture as Broxton but optimized for the Celeron/Pentium range of products... However, with Broxton out of the picture, it means Apollo Lake can be applied to devices it normally would not have been considered for before and Intel could release lower end versions that would have otherwise fallen into the Broxton category if there's still enough demand... at least for a tablet product.

But without subsidizing, it will rely on its naturally lower costs to achieve affordable pricing... Meaning, aside from products still being made on Bay Trail and Cherry Trail, we won't be seeing any new cheap $99 tablets... at best we may see something above $150...

On the plus side, Apollo Lake still offers the performance gains we expected from Broxton and it will offer OEMs up to 7% lower BOM costs, with reduced space requirements that allows it to serve a larger range of form factors and still will support mobile hardware components for low cost and low power consumption configurations more ideal for mobile tablets.

The Goldmont architecture is also more flexible, customizable, and optimizable than previous ATOM architecture... So there's some potential even if Intel never applies it to the phone market.

Exact figures have not been released on performance and configuration range of models that will be offered at this time but we should hear more by August...

Intel hasn't given up on mobile entirely, however, as they will focus now on 5G support going forward... This will primarily mean things like offering 5G modems and similar components instead of their own SoCs but Intel is still on track to reach 7nm FAB that will also switch away from Silicon to some other material that may help lower costs and ensure Moore's law can still be followed for a bit longer...

If they achieve this then they may revisit the mobile market but that will be a few years from now... So mainly only look for new Intel products in 2 in 1's on up, as well as Cloudbooks, Chromebooks, and other cheap laptops will also still be supported...

It's just that the ATOM branding may no longer be used as Intel has already started using other branding for the non-mobile product ranges... The Budget/Value laptop range is covered by the Celeron/Pentium Branding, for example.

It's also likely Intel will completely redo the architecture going forward, so even if they return to mobile devices it's likely they won't be called ATOMs anymore...

So, even though the ATOM is still technically around, the ending of the branding for the ATOM may be the permanent change... Aside from the older products that will continue to be sold for now for the remainder of their product life cycle.

Meaning, on one side we won't see a immediate change... It will be just that the Broxton and SoFIA updates will never be released, but as they become increasingly obsolete we will likely see ARM start to completely dominate the mobile market... Leaving just 2 in 1's and some higher end tablets with Intel inside...

Also, another range that's still up in the air is the IoT market... While ATOM based, they're using lower range Quark processors and it remains to be seen if Intel will update them but there's less need to update as quickly in the IoT range and that gives them time, at least for that part of the embedded market... as well as the Server market may still be served with ATOM based solutions, as this change mainly just effects the general consumer market.

Things to consider, is Intel may just merge what's left of the ATOM with the Core M to produce a more affordable, scalable, and customizable solution that could better leverage their technology to their advantage... Perhaps provide a hardware equivalent to what MS is doing with its Universal App Platform. So the same SoC can serve multiple duties and scale to the user's needs...

Imagine a hand held device that can not only change form factor when docked but also scale performance to that form factor... Such a game changer may be the end result of these changes... We'll see, it's just going to be a long couple of years until these changes come to pass...

In the meantime, this may promote Android and even iOS over Windows on mobile devices... as MS Windows 10 Mobile is unlikely to grow much for the next year or two and without Continuum and Universal App ecosystem it's likely we'll see fewer desktop OS products in the mobile range... albeit, except for the 2 in 1 part of the market that's fast becoming the new sweet spot...