Intel has been running its foundries to the limit since the past year and if what we are hearing is to be believed than the company is planning to offload some of its low end CPUs to GloFo. Keep in mind this is very significant news and since only one of our sources mentioned this and we have no way (as of now) to vet this, we will be marking this as a rumor instead of an exclusive. So take this at face value and keep those salt shakers handy.
Rumor: Intel will be offloading select CPUs to GloFo in 2020
That Intel underestimated demand for its 14nm process is an understatement. Having just knocked off a record quarter the company is already trying to increase fab capacity by 25% but even that might not be enough to survive the onslaught of expected growth of the PC industry in 2020. Since Intel's GPUs are also supposed to enter the fray in terms of production this year, things are going to be very tight and the company needs to manage its production lines very carefully in order to have a chance of meeting demand and not bleeding more market/mind share to AMD. This is where this particular rumor comes in.
Since GloFo is stuck at the 14nm/16nm process for now this would make sense as Intel would be freeing up critical fab space for its higher end processors while moving these parts where the market is fairly stagnant to GloFo. Another reason why this rumor makes sense is that the transition to GloFo is in line with their vision (as opposed to TSMC) considering the former is located inside the United States and the status quo socio-political scenario right now wouldn't look too kindly at outsourcing capability to TSMC.
According to our source, here are the lineups Intel is considering moving to GloFo:
- Celerons (very likely)
- Pentiums (very likely)
- Core i3s (maybe)
There have also been rumors that Intel is considering tapping TSMC for its next-generation graphics cards (certainly not the DG1) which would free up a ton of space at the cost of relying on an outside foundry for production. While I do not believe it is in Intel's best interests, this strategy also paves the way for a future spin off of the manufacturing group and allowing the company to become independent as a design house (that said, AMD's Wafer Supply Agreement is essentially a dead weight and while the company has managed to creatively work around it, the same might not be true for a post-spun-off Intel).
The choice to offload Celerons and Pentiums makes the most sense because these are processors which don't really need the cutting edge lithography or node maturity that is available in Intel's 14nm+++. GlobalFoundries nodes would suffice just as much in these lower tier platforms and I think it makes definitive business sense to get rid of these. The Core i3 part on the other hand I am not so sure off because that is a mainstream desktop part and one that is closely tied to the Intel brand name and since GloFo's process is going to be different to Intel's (unless the company shares its trade secret 'CopyExactly") it could potentially damage the company's goodwill.
All that said, if this turns out to be true, Intel should be able to add a few percentage points of capacity on top of the 25% this quarter and if the ramp to 10nm goes successfully would be finally out of murky waters by early 2021. That this is going to be a make or break time for the company would also be an understatement. If 10nm fails, then the company is looking at loosing a lot more market-share to AMD and once the 14nm node becomes old enough, demand would quickly fall as well.
