With all due respect, "never" is a real long time. Actually, the 10nm fiasco really wasn't that long ago and before that Intel was at least head-to-head with TSMC, if not ahead.
Intel Corporation to Participate in Upcoming Investor Conferences
- Thread starter XYang2023
- Start date
That's not what you said earlier:
"Yes, this is the same thing AVGO has done with every company they have acquired. All acquisitions must hit the target margins. If not, prices rise and/or heads roll until margins are met. If still not met, the group is sold or otherwise discarded. Investors love this scheme, other stakeholders not nearly so much.
For each acquisition this juice squeezing works for about 5 years at the max. So AVGO is always lining up the next trophy prey to pull into the processing plant. About a 2 year cadence iirc."
siliconbruh999
Well-known member
Before 10nm Intel was definitely ahead up until 14nm
My key issue with Broadcom acquiring Intel is cancelling of many Open Initiatives at Intel something that has drivern the industry Broadcom with Intel acquisition with IMS Nano fabrication they can even squeeze TSMC jump to Intels fab in long term and almost squeeze any Company with Intels assets
Last edited:
Just finished listening to the conference:
www.intc.com
MJ appears to be highly competitive and pragmatic, which instills confidence in Intel's products.
Barclays 22nd Annual Global Technology Conference
www.intc.com
MJ appears to be highly competitive and pragmatic, which instills confidence in Intel's products.
I believe the board's decision to appoint MJ as the CEO of Intel Products was the right move. She mentioned Falcon Shores, the next data center GPU. Her objective is to launch it, gather customer feedback, and quickly iterate on it. This approach aligns with my earlier opinion that Intel should not have canceled Rialto Bridge. It should be part of the learning process, much like what was done with ARC Alchemist.
siliconbruh999
Well-known member
Those decisions were already taken by Pat but he didn't know how to publicise them properly mj is handling it well i don't think she has Pat's technical expertise
I don't think expertise is necessary in this case. The reason is that MJ is already surrounded by experts. Listening to customers and understanding the market is more important. Based on what I’ve read, it seems that Pat was the one who decided to cancel the Riato Bridge, and he also intended to cancel the desktop GPUs. At times, I’ve wondered whether he paid attention to the broader situation beyond the foundry. I believe desktop GPUs are important because they allow students to purchase them and become familiar with oneAPI, which could enable them to use Intel's data center GPUs later in their careers.
I guess this is the part of my note you take issue with, based on your note about AVGO building AI ASICs for custoners.
Yep they have retained some folks who have figured out how to survive or maybe even flourish doing things the AVGO way. This benefits AVGO, of course.
But is this AVGO's main goal, or a secondary, nice to have outcome? In my view AVGO focuses on the next crop to juice (or herd to milk).
I guess we can agree to disagree.
siliconbruh999
Well-known member
He never cancelled Desktop dGPU otherwise we won't have battlemage
Rialto Bridge was axed by him sure
I meant that after the August earnings report, PG mentioned further 'portfolio trimming.' Then, there were rumors suggesting that dGPUs after Battlemage might be canceled. This was reflected in yesterday's reviews, with some reviewers casting doubt on the product's continuity.
Those “reviewers” know nothing of the roadmaps except what some random “leakers” say.
Tom Peterson in his official capacity has said Xe3 is hardware complete and in software development and Xe4 is in hardware development.
Note how I say Xe3 and not “Arc Celestial” because the IP development is moving forward but how they develop product lines and SKUs may not be as simple as just having discrete consumer GPUs.
And of course Xe is going to be in Falcon Shores and whatever comes next.
I stand by never.
Given the history of TSMC I do not see how another pure play foundry could even come close to what TSMC has accomplished. We shall see with Rapidus. Do you really think they have a chance?
Intel and TSMC have never really been head-to-head. Remember, you cannot accurately compare Intel IDM to TSMC pure-play foundry. TSMC has to accommodate hundreds of customers, thousands of PDKs and IP. Intel IDM only has to do Intel. A big example is power. TSMC has been power focused ever since the SoC. Qualcomm worked closely with TSMC on the Snapdragon series in the early 2000s then MediaTek and Apple. The only leading edge SoC not at TSMC is Samsung's Exynos and it is not competitive.
I really think we need to give TSMC credit where credit is due. I also think it is important to put Intel's foundry efforts in proper perspective. Otherwise, expectations can be unrealistic, and failure will closely follow.
The stock just popped by 25% due to high tech development of custom silicon for AI. This could grow by 4X or higher over the next 3 years.
Do you really think this isn't part of Broadcom's strategy? (including acquisitions)
Do you really think it isn't their main goal to grow through high tech investments that drive this growth ?