Ex-CEO Pat Gelsinger was against splitting up the company
Pat Gelsinger, former CEO of Intel, spoke out clearly against splitting up the company back in 2022. His statements in an interview at the time shed light on the circumstances of his resignation and suggest that strategic differences with the Board of Directors may have played a key role.
https://www.igorslab.de/en/ex-ceo-pat-gelsinger-was-against-splitting-up-the-company/
Historically FAB side suffers while other side prospers
Split has been coming for a long time - whether or not upper management admits it.
Foundry lost money and product groups made money only because foundry did not charge what they should have been to the product groups all along.
Had foundry been charging what they should have, changes would have been made long ago. This is an issue with IDM's - too easy to move money around.
@4wqy Certainly the Product groups have all the current profit margin and will continue to do so till they are sold or shut down. They have minimal capital requirements and rest/lounge upon a legacy business, so the cash is going to flow.
My point is that Foundry can always move to new nodes and is not tied to any architecture. Existing and past losses were due to the incestuous/co-dependent/symbiotic relationship with Product groups, where the fabs were run as a loss leader, providing all the excess capacity and cost over-run needed so that a dominant x86 never missed a sale.
That was smart and the right thing to do when the company had 80% of the market, but the company failed to adjust strategy, thinking that it could take back AMD market share once the fabs (somehow) got back ahead of TSMC. There is no explaining stupid.
Foundry being independent breaks the incestuous relationship with Product groups, maybe the single smartest thing Pat did (although he failed to grasp/accept how important it was for Foundry to be a separate company).
My take on this is that Foundry will be here a decade from now, when x86 is fondly remembered the same as all those other legacy architectures. Oh, those were the days..
Welcome to 6 months ago. Anyone who has been paying attention could see this coming - even Stevie Wonder.
Before split happens do everything in your power do be on other side opposite the fab. Look at all the splits - AMD/GF, ibm giving fabs to GF. Look at which side is doing well and which is struggling. Fab side always gets raw deal. Avoid it at all costs
@2xyt Foundry has lost money due to extremely poor capital and operational cost control (or lack thereof)
Intel has a long and sad history of excess capacity, the root cause of which are unrealistic sales projections which lead to silly wafer start projections. The root cause of that is well-documented lying by senior executives to each other.
I expect you will now see Foundry on a mission to get profitable. Fab managers historically made every effort to over equip with tools and headcount, in order to prioritize Max Output, so the Product groups could get all sales possible.
Those days are gone, and good riddance. What you should now see is fewer tool installs (based on PROVEN demand, not some nonsense sales target), and thus lower headcount.
I think Foundry is what should be retained, and that all the soon to be obsolete x86 product divisions should be made efficient and marketable, then sold.
If they wait till ARM is taking major PC market share, then the residual value of CCG craters alongside DCG.
I suspect both Pat, the Board and ELT understood all of this and the plan all along was to make Foundry suitable for external customers, and (at least as an option) make Product groups ready to be sold.
Pat failed his #1job, which is to sell Intel to potential customers. He made Foundry AND Product group promises which the company could not deliver. None of the competition will use a Foundry which is connected to a Product group. Period.
The problem on the Product side is that x86 is being overtaken by ARM, plain and simple. Aside from the ease of CUDA and the challenge of Gaudi software, NO ONE will use an AI product tied to x86, when they can get ARM solutions from multiple vendors.
The company needs more time to deliver on both, which is why D-mb and D-mber have clearly stated that the strategy remains unchanged.
Intel started reporting Foundry revenue in Jan 2024.
That tells me they had already had the split conversations.
And January they started the ball rolling to do the split.
Intel is a giant slow moving ship. A big change like that isn't
going to be quick. If it's going to happen it'll take time.
Here is an unpopular opinion: Intel has nothing to gain by splitting off manufacturing. The only thing it has in its favor is its all-in-one game. Their manufacturing cannot survive on it own onm its current track, it's losing $Bs a quarter with not end in sight. If they would focus on perfecting that process before all the expansion, they have a shot. But splitting off will only lead to losing the manufacturing, which is a huge money maker if working right. They need to address their problems. Pat being gone is a great first step, but they need leadership, and reduction of a couple layers of middle management, so they can get things done.
Intel always make wrong decision because it has too many PhD dump a-s without taking any advance high level psychology decision making class in school.
Intel not splitting earlier is one of the reasons for the downfall. Pat was wrong not to split it on day one when he arrived. That's also one of the problems. Engineers making business decisions that little business knowledge and are only chasing shinny ba--s that consumers dont care about.
"Split" should have came many years ago... Intel has been headed downward for at least 10 years. Anyone with a decent level of awareness and business acumen knows this.
He is probably the ex-CEO of the highest profile ever.