IDM and IFS with 4 in 5 is false prophet
- IDM where Intel makes its own silicon can’t continue. Even if LTD ever fixes its broken culture and gets on track technology 4 in 5 it isn’t enough. Not enough volume to ever learn fast enough or pay back the RD costs to compete effectively against AMD in x86 or Nvidia and AMD in graphics or those three and Apple and others in other areas. Not enough diversity and volume to make cost effective fabs. Any question why Intel needs tens of billions of government handouts ?
- . Hey lets be a foundry so we can get volume and learning and it so win win at the leading node, seriously? After what happens with Canon, Ice on 10nm who the fu€k will invest their leading edge product on IFS and hundreds of millions of dollars to see it fail. Best example is Qualcomm experience in one foundry versus another. Choose wrongly and years of planning and business fail. Let’s buy Tower it can make Intel great, sorry wrong foundry and to small and brings little. They needed to buy UMC or GF, epic fail and now no money and no scale.
Must be so sad for the employees moral
Again look at From Savvy @1lio+1ldXXqcP
IFS will never happen and here is why:
Intel lacks the ecosystem of tools, libraries and partners to broadly compete vs TSMC. this is 30 years to fix.
Intel does not keep 4-5 generation of legacy nodes around like TSMC. This is a big, big deal. Customers want to know they can get parts for years to come. Intel doesn't have the capital or the low cost structure needed to keep 4 generations around.
Manufacturing acumen needed for IFS is low volume, high mix of SKUs. Intel is low mix, high volume on its products. Intel has no understanding or skillset on how to be customer responsive and do highly flexible production planning for a low volume, high mix scenario.
The culture is broken. At Intel TD designs the process and you get what they give you. They are rude and arrogant and don't have the time of day for any product that is not leading edge and high volume. This will not sit well with potential IFS customers and the relationships and culture will simply not be compatible.
Let's say I am wrong on all of the above... o.k. lets talk about over what period of time and at what cost would such a new venture take? I easily estimate that it would take Intel 10-20 years to invest to build the capability. The bill for that could be as much as $20 - $40 Billion. The problem is that the type of investor Intel has today will not stand for it. They will quickly grow impatient and cut the funding off. Already, Intel bonds have been downgraded (last week). I project in the next 18 months the dividend will be cut. The core business has degraded so far that Intel can no longer hope to fund the fantasy IFS ride with margin dollars from the core business
What do you think? Any comments on this analysis? What did I get wrong?