Thread regarding Intel Corp. layoffs

Foundry is FUBAR and will fail

In the ever-evolving semiconductor industry, companies must adapt and innovate to stay competitive. One such company attempting to navigate this challenging landscape is Intel. Once a dominant force in chip manufacturing, Intel has faced significant hurdles in its quest to establish a foothold in the semiconductor foundry business. In this blog post, we will delve into the insights shared by Dr. Yang Guang-Lei, a former Senior Technical Advisor at Intel and former Senior Director at TSMC, during his speech at SEMICON Taiwan. Dr. Yang highlights the inherent problems that have hindered Intel’s success in the foundry business

Lack of Customer-Centric Engineering at Intel:

Dr. Yang starts by emphasizing a critical issue at the core of Intel’s struggles – the lack of customer-centric engineering. While Intel possesses impressive semiconductor manufacturing capabilities and advanced process technologies, it has failed to prioritize customer needs effectively. In contrast, companies like TSMC have been successful in building strong relationships with customers and tailoring their offerings to meet specific requirements.

Dr. Yang recalls a personal experience with TSMC, where he traveled to the United States to engage with Qualcomm. This proactive approach resulted in Qualcomm choosing TSMC to manufacture its 65nm chips. In contrast, Intel’s more technology-driven approach often leaves customers feeling like they have little say in the process. This disconnect can lead to products that may be technically advanced but don’t align with what the market demands

Challenges for Intel :

Another significant challenge highlighted by Dr. Yang is the current semiconductor industry’s environmental landscape in the United States. He points out that the younger generation in the U.S. is predominantly inclined towards software-related careers, preferring companies like Microsoft, Google, and Meta. This shift has led to a shortage of talent in advanced semiconductor research and manufacturing.

Historically, the semiconductor industry in the U.S. heavily relied on immigrants, particularly from Asia, to fill crucial technical roles. However, with many of these talented individuals now returning to their home countries, the U.S. semiconductor industry faces a daunting task in maintaining its growth and competitiveness.

Skepticism Towards Government Initiatives:

In response to the U.S. government’s “Science and Chips Act,”, Dr. Yang expresses skepticism about its potential success. He points out that this amount pales in comparison to the annual revenue of a leading foundry like TSMC. Furthermore, when divided among numerous semiconductor companies, the individual funding amounts become relatively insignificant.

Dr. Yang also raises concerns about the sustainability of such government initiatives. Will they continue year after year, or is this just a short-term boost? Depending solely on government subsidies, he believes, is an unreliable path to industry success.

conclusion:

Intel’s ambition to enter the semiconductor foundry business faces significant challenges, as highlighted by Dr. Yang Guang-Lei. Addressing these challenges will require a fundamental shift in the company’s approach – one that places customers at the center of its engineering efforts. Additionally, navigating the evolving talent landscape and remaining skeptical but open to government initiatives will be crucial in determining Intel’s future success in the foundry business. In a rapidly changing industry, adaptability and customer-centricity remain key to survival.

https://techovedas.com/intels-foundry-is-doomed-to-fail-tsmc-director/

by
| 2127 views | | 10 replies (last September 12, 2023) | Reply
Post ID: @OP+1ox98luZ

10 replies (most recent on top)

It takes great wisdom to survive where they are. Contrast to Intel fabs. Haha.

by
| | Reply
Post ID: @2vny+1ox98luZ

Dr Yang has great wisdom. In addition, his colleagues Sum Ting Wong and Wi Tu Lo are equally astute.

by
| | Reply
Post ID: @2ksg+1ox98luZ

There is a thing called a virtuous cycle and when you got a business and system working and are in that cycle it’s hasta la vista baby for everyone else. Intel had that mojo for two decades and rode it to a decade of arrogant and id--tic leaders who fumbled mobile and AI.

Now TSMC has that virtuous cycle of scale, momentum, customers, ecosystems and culture base that is a moat Samsung and Intel can’t hope to compete. Maybe Xi and Biden can bring it all to an end and Intel can rise,but even all the government handouts without a war or blockade of Taiwan, Intel is done

by
| | Reply
Post ID: @1axf+1ox98luZ

Dr Yang has great wisdom. In addition, having been a customer of TSMC and worked at Intel, I would add:

  1. Intel has no ecosystem for IFS business. maybe this is part of what Dr Yang means by 'customer centric' but with Intel it is my way or the highway. TD/TMG is impossible to work with. They think then know everything and they have no idea to support wide range of flows
  1. cost structure. Intel cost structure is terrible. Intel over-engineers EVERYTHING. many layers of redundant process and people at Intel. Checkers checking the checkers. US based regulatory adds huge cost. from constructing a new fab to all of the overhead from EPA to SEC to EIEIO. US employees cost more and produce less. Intel has sever revenue per headcount problem compared to TSMC.
  1. Competing business are viewed as 'risk' by IFS customers. Intel has their own products they want to place into the fab. How can customers get guaranteed capacity? When push comes to shove, Intel does not have a flexible shuttle system. They cannot deal with many mature processes simultaneously. Intel cannot understand or deal with low volume, high mix of SKUs and customers.

Of course over time, it should be theoretically possible to fix any/all of this. However, the odds are low, the amount of time and money is much higher then Intel can spend before customers and investors grow very impatient and stop funding the boondoggle.

Would it be possible for an Airlines like United to suddenly decide they would like to become an airline maker? Well, United certainly knows something about planes... but, that is a far cry from United being able to beat Boeing or Airbus. Of course analogies are not perfect and this one isn't either... but just step back and think how different you have to operate when you are vertically integrated vs. 100% focused as a core supplier.

by
| | Reply
Post ID: @fei+1ox98luZ

One fundamental difference between TSMC, Samsung, and Intel is culture. A corporate’s culture is just the reflection of the societal culture of its leaders and majority of employees. East Asian culture, based on Confucianism, values doing more than talking, promotes meritocracy, and denounces nepotism. Intel’s dominating cultures, be it woke culture or other foreign cultures, value talking more than doing, abandon meritocracy, and embrace nepotism. That’s why Intel is performing so poorly. That’s why talented engineers are fleeing from Intel to east Asia.

by
| | Reply
Post ID: @zlg+1ox98luZ

@ylt.. Not true. Most top technologists in Asian semiconductor companies are PhDs from US universities. A combination of sh---y hiring practices (D&I nonsense, ITAR etc) by American firms, broken immigration system and deteriorating quality of life resulted in these engineers/graduates heading back to their country.

Chinese semiconductor firms like YMTC and SMIC are offering > 2.5X compensation as compared to American companies these days. This is effectively 10X increase in quality of life if you consider the cost of living offset.

https://www.reuters.com/technology/china-quietly-recruits-overseas-chip-talent-us-tightens-curbs-2023-08-24/

by
| | Reply
Post ID: @ldr+1ox98luZ

I do not agree that IFS would be FUBAR. It's true that we can not beat TSMC but we can still hang around in the black as a foundry. Biggest challenges are cultural and behavioral, can we make an about turn in them is the question.

by
| | Reply
Post ID: @bjg+1ox98luZ

Technology driven and customer driven are not mutually exclusive. The problem is forcing a poor technology on customers.

by
| | Reply
Post ID: @wtv+1ox98luZ

US universities don't crank out process engineers as they used to do. Leading edge research in chip technology is prohibitively expensive, well beyond the resources of the universities. Publishing is difficult, so the whole sector gradually perishes.

by
| | Reply
Post ID: @ylt+1ox98luZ

Completely agreed !
Mediocre people who lead Intel’s Government Affairs Team and who next to do nothing about the semi conductors sector (Bruce Andrews in the US and Hendrik Bourgeos in the EU) have been patting themselves on the pack for scoring a minor open goal in the form of the Chips Act. Because Pat G understands nothing about politics he thinks it’s great - but it has little relevance and is knotted in conditions and ifs and buts … the idea that it will save Intel is laughable and the fact that Intel believes it will is even more laughable.

by
| | Reply
Post ID: @zco+1ox98luZ

Post a reply

: