Thread by @IanCutress

Now I've had some time for the @Intel announcements yesterday to sink in, here's my brain dump.

A number of analysts were quite reserved when @PGelsinger joined Intel, stating that no matter what he did, we wouldn't see it for years. I said straight away that Pat can steer the intent of the company as soon as he sat in the seat. Today is a clear message

That being said, Intel last week is still the same as Intel today. Saying stuff doesn't mean much unless Intel does pivot, and it will take a few years to enable that pivot, but there will be a strong undercurrent of things to come throughout, with a continuous focus on 2023

Questions about how IFS will take shape, the IP licensing model as well as customer access, and in what time frame, all TBA. If IFS can execute with first non-Intel IP customers by early 2022, and in volume by mid-late 2022, then that's probably a win

Saying $20b for two new fabs is good, however it's only increasing the number of fabs on the Ocotillo site by 50% - they already have four fabs there. There's no word about wafer starts per month, although

Gelsinger does want to be in a position where a customer can order 5k wafers/week and Intel can supply. The fabs will be for IFS and Intel's own use, and questions on which side has priority will be interesting. Could do the Samsung model here

How exactly Intel plans to expand other facilities, or create new ones, across the US and EU over the next decade will come into play, but no hard data on that yet. Also how much of this is available to IFS is a key question.

Tbh, the collab with IBM is going to be a big deal. Intel did not go into details, which is a shame. Every semicon conference I go to is filled with IBM papers and designs about current nodes and upcoming nodes and how to get different performance based on what you want

IBM used to have an outlet for this research, now of course they're relying on some licensing format to derive revenue. An IBM collab could be as simple as a patent licensing, or/plus a cross agreement that Intel might create a special node for HP (like GF used to do)

Also anything co-designed, given that DTCO is a big deal these days. No doubt if HP gets some custom access, they'll help DTCO Intel's own designs (either on first or third-party foundry) to the max.

In the past, Intel's main manufacturing was purely for its leading-edge CPU design. Now fabs are more nuanced, for RF, analog, high voltage, MEMs, etc. TSMC has roadmaps for these. If IFS wants to be universal, for Arm/RISCV too, it has to expand capabilities

IFS being its own business unit will likely mean that Intel will be enabling more outreach about its IFS services. Heading IFS is Dr. Randhir Thakur, who spent 14 years at Applied Materials working in fab operations, and understands supply chain, and has 300 EE patents.

If Dr. Thakur is available for an interview, then there are a number of good questions to ask :)