I really hope this doesn't affect the factory technicians again. They are already at skeleton crew numbers and struggling to keep up with the holes the last wave of layoffs left in staffing.
15 replies (most recent on top)
@fn, “ If anything the workload will get worse over time, and that is also more like a typical foundry operating model. But at least the tools will no longer be treated like little princesses.”
I agree with that statement too and it’s okay with the workload getting worse as long as management understands the priorities and if you have the right techs with the right knowledge and abilities across all shifts, then you can maintain a fair amount of equipment with a lower tech headcount. Right now that is not always the case some areas lost most of their experience in the last VSP, which did leave big holes like the OP stated, plus not all techs are the same, trying to put it nicely.
One thing I don’t think Intel understands or maybe they do but they don’t care is that it takes probably up to a year for a tech to become proficient at their job or at least become some value added, you can’t really just plug and play. If Intel thinks they will fill the holes with CW’s there are a lot of misses with those workers.
@fs+1 I always felt the change should have happened with IDM 2.0 and the separation of IFS and Products.
Products got the message but the slug Keyvan did basically nothing but continue to build out even more excess capacity. That was definitely IDM 1.0 thinking, and why many feel the only way to fix IFS is to spin it off or sell it outright, because the entire management structure likely needs to be replaced in order to make the changes needed.
LBT may be in a position to make that happen in-situ, even if he ultimately spins off, breaks up or otherwise sells off IFS. I think some combination of all these changes is going to happen, and that the effort is already underway.
@fn, Like I said before I agree with you, the priorities and how the factories are ran and structured needs to change. The factories have a lot of issues most of which are not related to tech headcount. Intel cannot run a foundry like it ran the factory for its own products, it’s a whole different mindset and that mindset might be hard to change, time will tell.
@fk+1 That is because they have not yet changed the priorities, but as you can see with the steady decline of the capital forecast, the changes are coming.
It starts with designing the fabs with less tools (actually starts with not building excess shells), and that requires a shift in priorities away from what was done in the IDM model.
When the company dominated markets and had pricing power, they made a choice to maximize both output and yield. That is really expensive and not how any foundry operates. It means not only excess tools but also excess maintenance. Foundries operate at higher capacity utilization by relying on run-to-fail maintenance and only having capacity which is supported by customer contracts.
IDM had no such contracts, and the roadmaps they produced (which was the IDM form of contract) were often fantasy and so changed every few weeks. Just producing that nonsense took over a thousand people.
There are far more issues than I'm describing, and they are all known by fab management and IFS management. The intent is to move away from that operating model, because the company can't afford such cash burn.
If anything the workload will get worse over time, and that is also more like a typical foundry operating model. But at least the tools will no longer be treated like little princesses.
@fj, I don’t disagree with you on some of your points, but just like the OP said some factories are running on skeleton crews and that is across all groups. It’s becoming harder as a tech in the factory to maintain equipment for what process is currently running, what is up and coming and what is being converted for foundry, it is a juggling act every day. For people on the outside to say curtain things it tells me they have no clue.
@fh+1 The comments about fab benchmarking are because it has been known for over a decade that Intel fabs were still optimized for max output, at the expense of cost effectiveness.
You will soon see the difference and it is a stark contrast.
Later on AI and robotics will greatly reduce the need for factory headcount, as both technologies are virtually made for semi production, with both huge cost savings and increased velocity.
But for now what the fabs are dealing with is massive losses and a lot of that is due to the excess headcount they have always had. Some of that excess headcount was due to having too many tools installed, and the reason for that was, wait for it, Max Output.
When a fab is optimized to do hot lots, max yield, and other costly functions, that takes extra tools and that takes extra headcount. Likewise in departments outside the fab.
These functions existed because the internal customers didn't have to directly pay for them. Product groups are not asking TSMC for these services, because they can't afford it.
Change is coming..and fast.
It’s funny how employees outside the factories point their fingers at the factories and say they are the problem, they are over staffed, they take long lunches etc etc, but they really have no clue. Perhaps there might be some Intel factories that are over staffed but that’s not the case for all of them. They try and compare TSMC staffing versus Intel staffing and yet they have never been to TSMC, they really don’t know the background, and therefore they really have no clue to that as well.
@aa, keep convincing yourself of that right up until you get laid off.
@ef+ It will take time but watch and see how IFS is restructured, to bring it to profitability regardless of external customer demand.
They held on to what they knew to be excess headcount, in the belief that they could ramp external customers faster than other foundries have ever done. D'oh!
Foundry management know exactly how they compare to benchmarks and what has to be done to get there. This has been examined in excruciating detail.
All this falls under restructuring so not a layoff which is announced.
In addition I expect to see most or all of the non-EUV fabs either sold or transferred to joint ventures with private equity or an existing foundry. I think this is where the rumors of the GF CEO moving to Intel came from.
Once those fabs are no longer part of Intel, they would be restructured even more and faster than what Intel fabs will experience.
And none of this even accounts for how AI and humanoid robotics are going to transform semi manufacturing (and design). That is where the next 50% reduction comes from.
Intel has a factory? More like a museum of ancient tools.
Intel bloated manufacturing has way more tech per wafer produced then TSMC.
Do the benchmarking vs. competition and then cut out all the bloat.
"These actions are necessary, but may not be sufficient to get to profitability. You people have been warned."
I thought we got rid of all the "holes" already.
Oh well, time to lay off some more.
Only exempt do the 5 hour lunch. Yeah they get drunk at lunch also.
I don't believe OP realizes the fabs are where all the excess headcount resides.
Lol uh huh. Techs will still keep taking their 5 hour lunches, so no change