Intel: CPU Shortages Will Persist Throughout Q3 2019
by Anton Shilov on April 26, 2019 10:00 AM EST
Although Intel invested some additional $1.5 billion to boost its 14 nm fab output last year, it looks like its supply problems are not going to be solved until the second half of this year. The company admitted on Thursday during its earnings conference call that supply challenges will persist throughout the third quarter.
“We have increased capacity to improve our position in the second half, although product mix will continue to be a challenge in the third quarter as our teams align available supply with customer demand,” said Robert Swan, CEO of Intel.
Because of record demand for server and high-performance client processors last year, Intel faced difficulties meeting demand for these products in 2018. As a result, the company had to invest $1.5 billion in manufacturing tools to increase output of its CPUs and chipsets made using its 14 nm process technologies in Oregon, Arizona, Ireland, and Israel.
Due to obvious financial reasons, Intel prioritized production of high-profile products like Xeon or Core i7/i9 over manufacturing of lower-end products, such as Atom, Celeron, or Pentium. This practice is expected to continue, so it will be somewhat easier to obtain a high-end part rather than an entry-level processor at least until the end of Q3.
A bit of good news, however, is that Intel started production of its Ice Lake-U CPUs in the first quarter. And, because of its factory network optimization, the company can produce more of such processors than it initially anticipated. Volume production of 10 nm CPUs will reduce pressure on 14 nm capacity and to a degree lower demand for 14 nm mobile products. As a result, the supply situation with Intel’s products made using 14 nm process technologies will likely be generally better in the second half of 2019.
Related Reading:
- Intel Starts Qualification of Ice Lake CPUs, Raises 10nm Volume Expectation for 2019
- Intel Further Boosts CapEx to Meet Demand for 14nm Chips
- Intel Investing $1B to Meet 14nm Demand: Prioritizing High-End Core and Xeon
- Intel 10nm Production Update: Systems on Shelves For Holiday 2019
- Intel’s Xeon Scalable Roadmap Leaks: Cooper Lake-SP, Ice Lake-SP Due in 2020
Source: Intel
Facebook
Twitter
RSS
38 Comments
View All Comments
JoeyJoJo123 - Friday, April 26, 2019 - link
I'm unsure if iGPUs are really the cause of any delay here, so tossing them out might not even help the node shrink process.levizx - Saturday, April 27, 2019 - link
Have you even seen the dieshots? iGPU is up to half the die, if they cut the iGPU in half, that's instantly 20% more yield.GreenReaper - Tuesday, April 30, 2019 - link
Right, well, they've been selling those iGPU-less ones, maybe that helped them.RSAUser - Wednesday, May 1, 2019 - link
Not necessarily since could be that the iGPU is fault tolerant, while the CPU could maybe have a lower fault tolerance, so could just waste wafer space.DanNeely - Friday, April 26, 2019 - link
Because most of their consumer CPUs go to OEMs who want to build systems as cheaply as possible; which means they want an IGP so they don't need to source a separate graphics card.Mainstream enthusiast systems with discrete cards aren't a big enough seller to be worth the (very large) cost of creating an additional die variant.
bobhumplick - Friday, April 26, 2019 - link
actually im pretty sure they are gonna do just that with 2 lines of dies, one with igpu and one without. possibly just have no igpu on the die and add one on a seperate die.the current bios updates to the R0 stepping mean a new chip or chips is being made. while they are at it i think they are going to leave the igpu off at least one line of dies. they may already have enough to supply cpus that need an igpu. the current 9900kfp and 9700kf have the igpu turned off but its still wasted space. i think they did that to get people used to the idea and test market it (not to mention get more cpus out).
but think about this. the igpu on the current i7's take up about 3 cores worth of space. you could make a 9900k cpu with no igpu that would be between a 7700k (quad core) and an 8700k (6 core).
thats going to be the next i7 according to wikichpi with the next i5 being the 9700k basically wtih 8 cores 8 threads.
and wikichip has it listed as working on current z370 boards. not having to make as many new chipsets means less demand on the 14nm node so more cpus.
i think they can compete with that lineup (would like to see the upcoming 10 core 20 thread be the i7 instead of the i9 but...) next gen. after that who knows.
and heat and power wont be that bad for diy builders. any dual tower air cooler can handle the 10 core if you set the power limit according to what your cooler can handle.
my dh 15 can handle 220w very easily and keep a n 8700k delidded under 70c in just about anything but prime95 small ffts.
so you find the max oc you can get on the 10 core or 8 core. then you set a power limit according to your cooler size. at 220w a 10 core would probalby be able to run 4.6-4.7 all core in a rendering app and in gaming it will sustain a full 5ghz all core because game threads dont use much power.
plus no latency concerns with IF and multichip. amd will win hands down in pro work though
Gastec - Sunday, April 28, 2019 - link
Why isn't there even ONE capital letter in your comment?bananaforscale - Wednesday, May 1, 2019 - link
Oh but there are, in "R0" and "IF". :Dnandnandnand - Friday, April 26, 2019 - link
It looks like Gen11 graphics (64 execution units) and the Xe-based integrated graphics are going to be a huge improvement.I wouldn't count Intel out. Improving the integrated graphics is one of the easier problems they have.
HStewart - Thursday, May 2, 2019 - link
I think that is understatement - 14nm does matter to Intel anymore - so they don't care what is going to happen to supply 14nm - it is minimal market for Intel for game machines.Gen 11 graphics is replace Integrated graphics and supposedly have speed like 1030 or even higher and Xe is in total another league. I think Xe will eventually land it self on notebooks and have performance level of desktop chips.
But think about this is in another way - Intel has been bashed so much about 10nm and its 14nm - with billions of dollars and all the new hire smart personal, do you really think that Intel is just sitting around on behinds letting that happen. I believe they have something big coming and there is a lot rumors but they may have let them out to mislead the industry.
Now there are other signs that something big is coming, Intel has massive movement on their fabs like getting rid of older Micron fab. Updating fabs for news.
I would not count Intel out but keep in mind AMD is not Intel biggest threat in the long run, ARM cpu's are the bigger threat and could hit Intel where it hurts the most - so it very smart of releasing them first. But no one truly knows the performance of them and rumors are likely no where close. Also with new technology chips and changes in architexture, it very possible there will be no difference between desktop chips and mobile chips. This could be a high end laptop with Sunny Cove could performed the same as desktop.
Intel is not stupid, but above is what I would do if I was Intel. But time will tell.