This post was going to be about Intel cancelling the Lakefield chipset due to low demand and shifts in the market but I want to expand o that.
2020 created an interesting shift in computing. For a whole year computing shifted back to a more PC centric stance form mobile. I don’t think its a lasting trend but I think the shift in focus will have lasting impact on the PC market; and one of those areas is around adoption of ARM.
For most people ARM is the chips found in their smartphones and beyond that its stuff for the nerds. A few will remember the days ARM, and the similar RISC, powering early computers like the Acorn. It’s combination of performance and battery management has made it the de-facto chip for mobile devices and the System on a chip design (SOC) has allowed it in recent years to move into more spaces like the data center and PCs.
And this is the part where Intel comes in.
I don’t know how many people remember Moorestown or MeeGO or AppUp or any of the earlier initiatives Intel had in getting into the mobile market. Much of the early to mid 2000s found Intel trying to compete in the lower end chip market to really mixed results that led to them cutting out of the mobile market in the 2010s. Then in 2019 Intel started discussing Lakefield.
Lakefield was Intel’s take on SOC design. For years Intel has been working on getting its chips to run better on the low end. This means balancing the power and performance its core chips are known for with designs that allowed for things like fan-less PCs. Lakefield was supposed to be the next step in allowing Intel’s chips to be used in things that look more like the iPad and less like netbooks. It would also help in competing against the likes of Qualcomm which were coming into the laptop space thanks to Windows on Arm (WoA). And Intel even designed concepts for Lakefield devices which clearly showed this chip was made for mobile computing. Both Samsung and Lenovo were coming out with devices running on the chip while Microsoft was using it to power both the new Surface Neo and Windows 10x.
All good right?
Well no. Intel announced on this week it was ending Lakefield mostly due to demands of the market and customers going for other chips. Also it may have to do with the fact that it has struggled with getting its chips passed the 10 nanometer mark.
Okay this piece is running long and Intel is only part of the story.
I am going to rush this a bit so apologies if I have left out all the juicy architecture bits (not Anandtech). Intel canceling Lakefield doesn’t stop the trend in adopting mobile chip design to larger computing devices. It only changes the major players.
Okay the M1 chip by Apple is a big deal and a drop in the bucket and a major shift in thinking all at the same time. It is a big deal because it is a mobile first chip designed by a major computer vendor that competes and at times beats laptop processors. It moves ARM and SOC chips beyond mobile phones. It is another signifier of the larger shifts in computing not needing certain overheads. But it is a drop in the computing bucket. it runs on the Mac and one iPad model; everything else is running on Intel and AMD or Qualcomm.
Don’t get me wrong the M1 and its successors are a big deal, but its going to get competition. it won’t one to one. Apple has been good at getting its developers to move with the platform; but what they’ve done is repeatable. Microsoft moved to the ARM architecture during Windows 8 and while its been slow has been evolving. Qualcomm has moved from using off the shelf chips to designs built for Windows and ChromeOS. The reception of the M1 is helping push ARM chips further into the overall PC market; Intel dropping Lakefield is a potential boon for Qualcomm especially for PC vendors who are looking towards more mobile form factors.
Okay that’s it I have written enough gobbly wobbly.