In AXG QGS this morning, RK blamed macroeconomy and CPU market share shrink for AXG deficit. In fact, the deficit is probably close to the money loss due to PVC poor execution and his pipe dream of entering HPC and AI markets. If he focused on first getting integrated GPU right and then penetrating discrete GPU from the low end, AXG would be much better off today. His PVC push is a disaster for AXG and Intel, new architecture, new process, and second rate engineers. Total FUBAR. He is purely a showboat, w/o business nor engineering sense. No wonder Lisa Su kicked him out of AMD.
20 replies (most recent on top)
Not one single executive will be let go involuntary. It will be all those low end managers and workforce. How do I know this, just look at how Pat took a all paid trip to Hawaii recently with some execs and Lisa P in AXG took her whole team for a weekend trip paid for by Intel.
@2kcj+1jwCtL89 has good points. RK himself is a AMD reject. All the recent hires from AMD are also questionable. Since AMD is doing well, better than Intel these days, if they were important contributors to AMD successful story, why would they come to Intel? Most likely they are not key persons in AMD, or worse, rejects like Raja. I see few few, if any, outside hires better than Intel good engineers.
All PEs across Intel need to really be scrutinized for their suitability for the roles. There are many duds.
when donkey's fly
The formula for Intel to succeed is to replicate its own success story, getting ahead in process technology first, and ramping the fabs with a couple high volume products, i.e. Intel = TSMC + Apple + Nvidia. It is a tall order, but possible, if we have the same engineers and managers in Noyce, Gordon, and Grove era. Unfortunately the company is now run/ruined by bozos.
Rk reminds me of visionary who sees visions. He can’t execute, he doesn’t allow his staff to even question him, he doesn’t stick to the agenda, and he doesn’t even know how to coordinate within the organization. In 5 years, there is enough proof of his inability.
He supposedly had the responsibility for scrutinizing hiring of the best of the best. He hasn’t hired one such person. In fact, there is a plethora of “talking heads “ like himself in gfx. No wonder not a single successful gfx project. In fact, it seems like he was planted at Intel for just destroying any chance of Intel success 😑
Good analysis, but where is the focus exactly? I mean what got cut? Still trying discrete graphics. Still doing PLD (programmable logic / under server)? Still trying AI inside. Seems like all over the map. IDM 2.0 is huge bet against arguable the best company around -- TSMC. So, is it words regarding cuts and focus... or did something actually get cut?
The IDM 2.0 is just a bridge too far for me. Intel core competency is very high performance / high MIPs per watt. Very custom process. Non standard tools flow. Latest node. TSMC covers wide range of technologies with standard design flow on EDA tools. For Intel to try and go for this is like trying to throw a 99 yard touchdown pass. Not possible. At best you throw for 60 and hope to god your receiver can get into the end zone. Fun to see it happen but, I wouldn't bet on it.
I remember a few years ago in an all hands, either BK or BS explained why Intel was investing in all the different areas including AI, Discrete GPU, etc. Their reason was all of them have potential to grow to a big market, so Intel had to place "bets" in them. That to me was the leaders having no clue where the industry is going. They get paid millions of dollars to be visionaries, not placing bets on the future. Even if you were to place bets, you can't bet everywhere. You need to focus on a few areas that have the best chance to succeed and have a strategy(i.e. hiring the right talent) to succeed. Entering discrete GPU market when NVIDIA and AMD were already dominant makes no sense unless Intel was going to outspend them to hire the best GPU engineers in the industry. Intel wasn't hiring the best GPU architects or engineers in the industry, so how exactly did they expect to catch up to 2 already dominant players? I think Pat is doing a better job by getting Intel to focus on fewer things. Some people might not agree with IDM 2.0, but at least he's going all in on something rather than trying and failing in everything.
" My God, why did they need to wait for revenue decline to do that? "
Good question. When companies are profitable, the execs want to grow the business (and power base) to keep the earnings up and to the right... So, they invest in questionable projects. Along the way, if the project goes south it is hard to ki-l it. Someone has to admit it is failing. Nobody likes to admit failure. Instead, execs tend to wait for the big macro problems. Then, you can cut the project and blame external forces. It's human nature, that's all.
Intel should have never hired Raja. Ok, hiring was a mistake. He already had a bad reputation. Everyone who knew him said he had execution problems. Ok, Intel hired him. Or Murthy is responsible for his hiring. Intel should have let him loose so many years with zero accountability. This just shows the problem with management. Even if there is a hiring mistake, there should be internal processes to boot him out because he didn’t bring a single marketable GPU since 2017 - 5 years.
But no, when recently intel promoted him to EVp, our chairman sent a tweet and said “well deserved” well deserved based on what. It is also totally inappropriate for the chairman to congratulate a company exec.
AXG is Raja and his friends. He promoted his TA, who was an apple sales rep before Raja made him his TA. Crazy.
BTW, in all recent communications, the management is saying that they are doing portfolio reviews and cutting unsuccessful programs. Layoffs are not peanut buttered across, it is done surgically - by removing these bad businesses. My God, why did they need to wait for revenue decline to do that? If those programs were unsuccessful why did they start them or why did they give a green light at the beginning of the year. If we hit our revenue targets, does it mean, they would have continued these unsuccessful programs.
AMD has apparently released some high end graphics cards with top of the line performance and very good price for that performance. Are they also on TSMC? Just wondering how they did that?
@1gse+1jwCtL89 it is same everywhere.
Check how Twitter board pushed Elon to buy them out and pay them the money. Now Elon has bought see how much havoc he has created with layoffs... upper management always keep themselves secure first with money and then play all the cards.
Tell us how many layoffs in AXG?
@1aez+1jwCtL89 has good points, but PVC is a diverge from existing Xe GPU architecture and design, sucking too much resource, expanding HC too fast, and with negative ROI. Nvidia is too entrenched in AI and HPC markets with CUDA ecosystem. Unless Intel is able to deliver more than 2x performance at the same price, which I doubt, it’s a hard sell. The decision of pursuing PVC shows that RK is a very phony guy, incapable of making sound business and technical judgement. He should step down.
Raja is just another Murthy. Talk. Get the money. Wreck the place and leave.
I hope you are right but, don't fall into the trap of,
"what we are thinking of is better then what they have... "
Agreed with getting rid of RK. Though I don’t think the GPUs are as bad you might think. The flex series is doing very well and beats comp. Yes SPR level of delays for PVC but that architecture is a tiled base design not monolithic like NVDA which is why hopper yields are so bad. It beats those NVDA cards too. SW and FW are kinda of mess but better then AMD
Purge G11+
Wow sobering summary. I wish I had a nickel for every dollar Intel ever spent on going after discrete graphics... from i740 to present day. Next BU meeting ask the finance team how much it totals up to, but if you do make, sure you are sitting down.
AXG management as a whole needs a overhaul, to many VPs in the same positions dropping the ball year over year but never touched during layoffs.