You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
At this stage, do you think the US government should cancel the grants given to Intel and the money to TSMC in exchange for expanding AZ even more? For example, the US govt allowed ASML to acquire SVG, the last remaining US based litho company in the early 2000s. Why not realize market reality and go all in on TSMC? If Intel shuts down the fabs, TSMC can hire the unemployed Intel engineers and technicians for 50-75 cents on the dollar so it would make US based manufacturing more cost effective as well.
Process engineering will still be done in Taiwan. And TSMC is not putting their latest process anywhere else but in Taiwan. The Arizona fab will be one or two nodes behind the leading ones in Taiwan. Whoever thinks Apple will be making leading edge chips in Arizona is delusional.
As one of TSMC’s advanced fabs, TSMC Arizona will play a vital role in the U.S. government’s goal to onshore semiconductor manufacturing and strengthen national economic competitiveness. TSMC Arizona’s first fab will operate it’s leading-edge semiconductor process technology (N4 process), starting. production in the first half of 2025. The second fab will utilize its leading edge N3 and N2 process technology and be operational in 2028. The recently announced third fab will manufacture chips using 2nm or even more advanced process technology, with production starting by the end of the decade.
By 2025, when the first N4 fab is operational in Arizona, TSMC is supposed to be in volume production of N2 in Taiwan. N3 has been in volume production there for almost two years.
So I would say the US government has little choice but to continue funding Intel.
Process engineering will still be done in Taiwan. And TSMC is not putting their latest process anywhere else but in Taiwan. The Arizona fab will be one or two nodes behind the leading ones in Taiwan. Whoever thinks Apple will be making leading edge chips in Arizona is delusional.
By 2025, when the first N4 fab is operational in Arizona, TSMC is supposed to be in volume production of N2 in Taiwan. N3 has been in volume production there for almost two years.
So I would say the US government has little choice but to continue funding Intel.
You are correct about TSMC doing process development in Taiwan. However, there are only two proven foundry process development teams in the world, one in Taiwan and the other in South Korea. Given how hard process development is, isn’t it a better use of taxpayer money to try to secure friendly relations with TSMC? The US government does not have any issues with having litho development in the Netherlands and Japan. What’s wrong with having process development in allied Asian countries?
War with China will be so disastrous that it’s better to ensure that it does not happen. Therefore, I think geographic proximity to China at the end of the day is a nonissue. Yes, it would have been nice to have development capacity in the US but if talent is lacking, it’s best to face reality and deal with it by ensuring good ties with TSMC, Samsung and governments in the area.
War with China will be so disastrous that it’s better to ensure that it does not happen. Therefore, I think geographic proximity to China at the end of the day is a nonissue. Yes, it would have been nice to have development capacity in the US but if talent is lacking, it’s best to face reality and deal with it by ensuring good ties with TSMC, Samsung and governments in the area.
Unfortunately, there was not that many confrontations between the US and other countries that did not end up in a war. Saying that development capacity in US is lacking is... strange. There are plenty of technologies where US is ahead of everybody. Semiconductor technology is not one of them right now because for a long time outsourcing has been a cheaper option. But now, the cost becomes secondary to security. This changes the equation.
There are plenty of technologies where US is ahead of everybody. Semiconductor technology is not one of them right now because for a long time outsourcing has been a cheaper option. But now, the cost becomes secondary to security. This changes the equation.
The ignorance of this statement is simply astounding. AMAT, LAM, TEL, KLA, Cymer, Inpria, and many others I have never heard of do tool, material, and process research in the US. Every company on that list besides TEL is also HQd in the US. Even if you want to pretend that intel doesn't make "advanced logic", Micron is also the process leader in DRAM and in many aspects is also the leader in NAND. All of their pathfinding and early development happens in Boise ID. Cheaper labor also has no impact on where process R&D happens. If Asia is "just better" why has every Asian firm besides TSMC and Samsung fallen out of the advanced process race before intel or even GF for that matter? Why has intel passed Samsung? Why does Micron lead SK, Samsung, and Kioxia? Why are there no tier 1 Taiwanese or Korean tool vendors, and among the T1 vendors why are so few of them Japanese? TSMC's success has everything to do with TSMC and very little to do with Taiwan being magically better at semiconductors. To claim otherwise is to do a great disservice to the excellent work TSMC's engineers and technicians do.
Also worth pointing out that within a year or two the US will have leading edge Micron, intel, TSMC, and Samsung wafer fabs. The ROK only has Samsung/SK, the ROC only has TSMC/Micron, and Japan only has Kioxia. Combine this with the literary of trailing edge stuff scattered around the US and if anything the US has the most divesifed semiconductor industry of any one nation. While other nations at most only have 2 leading edge fab operators in their nations.
I'm not at all surprised by Intel's troubles. After 35+ years in the semiconductor industry, I reluctantly joined Intel as part of an acquisition in 2014 and spent about 5 years there. What I observed was arrogant leadership that was in complete denial about the competitiveness of their leading processes, discounted the threat of ARM, was extremely fat and inefficient, ridiculously bureaucratic, and operated in an inbred bubble. The issues that led to Pat's "Key Priorities" were very evident to me ten years ago. His current objectives seem to be on target, but the question is whether or not the Intel culture is willing and able to adapt before it's too late.
I'm not at all surprised by Intel's troubles. After 35+ years in the semiconductor industry, I reluctantly joined Intel as part of an acquisition in 2014 and spent about 5 years there. What I observed was arrogant leadership that was in complete denial about the competitiveness of their leading processes, discounted the threat of ARM, was extremely fat and inefficient, ridiculously bureaucratic, and operated in an inbred bubble. The issues that led to Pat's "Key Priorities" were very evident to me ten years ago. His current objectives seem to be on target, but the question is whether or not the Intel culture is willing and able to adapt before it's too late.
That is Something only time has answer to but they should let go of their leadership mindset and accept they are not what they used to be only than something can happen else it is gonna be same story
A little OT — is Intel unique in offering sabbaticals for longer term employees in this industry or is that more common?
I’ve managed / directed a lot of very senior engineers and coached/approved many of them on how to work the system to take medium term sabbaticals (i.e. very long vacations), but the (very large) employers I’ve worked for never had anything formal in this area. I’ve known a lot of employees who felt (esp after many years of service) that the few weeks of vacation wasn’t enough to let them explore things especially as half of that usually goes to short term / family obligations.
Because Intel eliminated the free fruit program for the office, will it be better to bring fruit to them instead of (or in addition to) pizzas?
Intel paid big EDA vendors to port their tools to Intel Foundry. They even started a $1 billion fund to grow the Intel Foundry ecosystem. Although we don't know how much money left after the $10 billion expense cut announced.
That is Something only time has answer to but they should let go of their leadership mindset and accept they are not what they used to be only than something can happen else it is gonna be same story
One of the things I have said over and over on this form is that Intel still see's itself as the 800lb gorilla of semi conductors, when the reality is today they are just a chimp. As long as they fail to acknowledge the reality, they will continue to blunder.
Much of Intel's current conundrum stems from the IDM-knows-all way they worked with fab equipment, EDA and IP vendors in the past. Intel explained how fab tools, EDA algorithms and flows, plus IP must work to the suppliers to support Intel's magically powerful process and design, instead of Intel leveraging the suppliers to figure out the most efficient and effective way to get the end results they wanted to achieve. From everything I see, Pat is transforming that part of their culture, enabling Intel to learn from the outside. One of the biggest benefits from smaller EDA companies is the ability to adapt their EDA product(s) and associated innovation to a specific customer's needs, but that's kind of the opposite of what Intel needs right now - they need to be able to to leverage the productivity and efficiency capabilities EDA and IP suppliers have delivered for competitors like TSMC, AMD and NVIDIA.
From the Intel company website:
...
Next week, we’ll announce a companywide enhanced retirement offering for eligible employees and broadly offer an application program for voluntary departures. I believe that how we implement these changes is just as important as the changes themselves, and we will adhere to Intel values throughout this process.
I participated in a voluntary departure program once. Not in semiconductor industry, it was the IT organisation of an international agency. In that case, most of the best people left. I hope it will be different for Intel, but the incentives of the voluntary departure program are backwards: who can find new jobs has an incentive to leave, who can't has an incentive to stay
WASHINGTON—August 5, 2024—The Semiconductor Industry Association (SIA) today announced global semiconductor industry sales totaled $149.9 billion during the second quarter of 2024, an increase of 18.3% compared to the second quarter of 2023 and 6.5% more than the first quarter of 2024. Sales in...
I participated in a voluntary departure program once. Not in semiconductor industry, it was the IT organisation of an international agency. In that case, most of the best people left. I hope it will be different for Intel, but the incentives of the voluntary departure program are backwards: who can find new jobs has an incentive to leave, who can't has an incentive to stay
Yeah, I have seen the same thing happen at another company I worked at that had a voluntary departure. The company was shocked when so many high performers left (I am not sure what they were expecting - that the worst performers would self select out of the company?). Some teams lost 20% of their people, and mostly the best people.
