IDM 1.0 Product groups benefited from having access to lead nodes (ahead of what other companies could access), to the point that the fabs were run as a loss-leader, with max emphasis on output and yield over cost.
This led Finance to grind their teeth but no one cares about that.
It seems clear that external customers want Floundry to be a separate company from Products, because of the concern over IP sharing and wafer start conflicts.
But that leaves Product groups with low margins when they rely on lead nodes from TSMC. This is because they no longer have pricing power and that will only worsen over time as x86 is supplanted by ARM.
So Product groups NOW need Floundry to be a more cost effective supplier TSMC (and close to leading node). This was the point of IDM 2.0.
MJ tried to mumble something about this in a more positive light, but the reality seems clear enough.
What this likely means is deep cuts in Floundry spending, as that group gets real about the capacity needed for the pace of customer onboarding which is possible.
The pushout of Ohio and halting of other projects shows the effort is underway to rationalize capacity to demand.
Pat was pushing to do a full buildout, which only made sense if he was able to bring high volume customers onboard.
A smarter approach would have been to do no greenfield projects.
Just add a few mods and wait for customer growth to justify Ohio and Germany.
This is what appears to be the current plan (much to Pats deep chagrin)
Next is to slow down the ramp, and stop building speculative capacity.
The company has being driven into the ground by reckless expansion and it must stop.
So at the existing facilities, that means fewer tools, which means less headcount. Attrition may be sufficient.
It seems possible that a few HVM fabs could be spun off into an independent company, to satisfy external customer concerns.
Considering how few customers there are, that could just be at one site, like Ireland or something.
For Product groups, it is way past time to stop projects and groups that have no roadmap to profitability.
Because that x86 market share, it ain't coming back.
If Product groups need the combined margins then they will have to retain some fabs in the same company, and likely TD as well, in order for it to be funded.
So the big surprise of 2025 may be that the company is not particularly broken up, but that the fabs are broken up, in order to sustainably serve the needs of internal and external customers.