poke01
Diamond Member
- Mar 8, 2022
- 3,309
- 4,559
- 106
Can’t argue with that,Of all the shades of green, Nvidia is the slimiest
Can’t argue with that,Of all the shades of green, Nvidia is the slimiest
Intel announces Lip-Bu Tan as new CEO, effective March 18th.
It still is a change of staff since UC is basically handing core development off to Intel Austin.
My guess is that LBTs mandate is to stop the bleeding with foundry 2.0. He knows Pat’s plan and he doesn’t really have time to change much other than dramatically trim the work force. Intel has $50 billion in foundry spending that they are bringing on line.Very funny development since the BoD didn't listen to him while he was part of it, so what will be the change when he's CEO? The BoD ended up firing Pat without base in reality, it will eventually do the same with LBT. Let's see what he manages to change until then (changes visible right away since we can't expect the Intel BoD to be able to look further than one quarter).
I think things collapsed faster than Pat expected and he made some bad bets on wafer demand with Apollo and Brookfield. There were already late penalties last quarter for Ireland. Skipping 20A might mean there are more penalties in Arizona. There is a lot of capital that needs to be deployed. LBT needs to get fabs up and running full with x86 customers or anybody with money.
What AI biz?GFX division is still necessary for competitive IP for SoCs and to scale out AI biz. I think discrete efforts will get the axe but not the division as a whole.
Intel had to settle for poor man's Tan? Kind of like a spray tan?
After Hock Tan of Broadcom said "no"?
A tragedy for the ages. Just think of the influence we could have had! Finally, SMT4 cores! Just not at AMD.They went with him? Damn, they must not have received my application!
Opinions can vary WILDLY. I'm pretty sure there are others that vehemently disagree with you.AMD cornered desktop DIY with something between zero (up to Zen3) and "barely existing" (Zen4/5) IGPs, all those 32EU IGPs on faster desktop CPUs are objectively wasted margin for the most part.
Intel should've long ago started doing a mobile-focused 6+8+ 32 EU and a desktop-focused 8+16+ 8 EU, as the latter will usually be paired with discrete GFX from AMD or NV anyway.
The whole "stop the leak and stabilize the ship and hand it off to a proper CEO" was supposedly Bob Swan --> Pat Gelsinger.Makes sense. Hire a person whose job is to stop the leak and stabilize the ship, then hand it off to a younger CEO who can take on more risk and make big bets. LBT is 65 years old. I highly doubt he'll be the Lisa Su of Intel.
Also, they shouldn't abandon the dGPU market. Half of their entire problem is giving up on a project too early. If you are going to do that, DON'T START AT ALL!
Lip-Bu has strong views on AI business and the fact Intel didn't had anything competitive or a solid roadmap to tackle the market was one of the reason he disagreed with the previous board.What AI biz?
On Desktop I do agree. Intel ships too much GFX silicon compared to AMD.Intel should've long ago started doing a mobile-focused 6+8+ 32 EU and a desktop-focused 8+16+ 8 EU, as the latter will usually be paired with discrete GFX from AMD or NV anyway.
Lip-Bu has strong views on AI business and the fact Intel didn't had anything competitive or a solid roadmap to tackle the market was one of the reason he disagreed with the previous board.
Did they? Outside of Gaudi, which Intel killed in another series of stupid decisions, what else did they had?Intel had products that were in theory competitive performance wise. Nobody cared.
If it follows AMD recipe it should be cost efficient when it comes to the fixed development cost.16+32 is dead for sure.
Knights Mill? True, it didn't quite have AMX, just souped up AVX-512.What's really funny is that Intel could've created a Xeon Phi type of product with AMX instruction set pretty quickly for AI workloads. Instead, they kept dancing around, like a kid in a candy store who has no idea what he wants.
People would go crazy for it, especially if it becomes sort of a middle ground between consumer dual channel and HEDT 8-channels with quad channel RAM support.Do you think it would be unprofitable due to lack of demand from customers?
Yeah, supposedly but the Intel that PG inherited wasn’t exactly a lean Intel, and then PG spent money like a drunk sailor which is kind of antithetical to the Rory Read to Lisa Su transition where AMD stayed lean and focused throughout her tenure. PG bet big and basically lost, and now they got LBT to come in, do damage control, and salvage whatever he can. I honestly think he has a larger hill to climb than PG when PG started his role, simply based on there not being Covid to boost chip demand, AMD/Nvidia are solidly more ahead than before, and employee morale is going to take an even bigger hit if he had to make cuts.The whole "stop the leak and stabilize the ship and hand it off to a proper CEO" was supposedly Bob Swan --> Pat Gelsinger.
IMO this is just a continuation of a broader trend which results in US ultimately losing the final area in technology leadership(and economic leadership). Just like the market there will be ups and downs but ultimately a decline.
Absolutely agree with you. Whatever good Pat did, it will be forgotten if LBT succeeds in steering Intel in the right direction despite the rather grim prospects for Intel's future at the moment. If that happens, I would then say Pat was necessary because most likely Intel would never have considered LBT for the CEO position had Pat not messed up so badly.I honestly think he has a larger hill to climb than PG when PG started his role, simply based on there not being Covid to boost chip demand, AMD/Nvidia are solidly more ahead than before, and employee morale is going to take an even bigger hit if he had to make cuts.
well Intel only has to stop being schizophrenic.Lip-Bu has strong views on AI business and the fact Intel didn't had anything competitive or a solid roadmap to tackle the market was one of the reason he disagreed with the previous board.
I'm sure all 12 of those customers would be ecstaticPeople would go crazy for it, especially if it becomes sort of a middle ground between consumer dual channel and HEDT 8-channels with quad channel RAM support.
Nah. There's lots of such people. I will state my own example. I think the world as a whole looks to be going towards increased instability and uncertainty in every sense of the word so I am finding it easier and easier to throw money away at some hardware to experience the good stuff before everything comes crashing down.I'm sure all 12 of those customers would be ecstatic