New Intel CEO
- Thread starter Artificer60
- Start date
1. IFS does make normal PDKs but they are not as good as TSMC nor will they ever be. TSMC has 30+ years experience making PDKs for customers and TSMC has the top semiconductor companies helping them.
2. Bribing EDA vendors is a VERY bad idea. There are no secrets in the semiconductor ecosystem.
3. See above and once you are on your knees with customers it is VERY hard to get back up.
The most important selling point for a foundry, especially a leading edge foundry, is trust. That should be the first thing IFS thinks about everyday, how do I win the trust of a foundry customer. If you don't have trust you will not succeed in the foundry business. Just ask Samsung.
I think at this moment, it is not just trust. TSMC holds a monopoly position. Obviously fabless companies would use them.1. IFS does make normal PDKs but they are not as good as TSMC nor will they ever be. TSMC has 30+ years experience making PDKs for customers and TSMC has the top semiconductor companies helping them.
2. Bribing EDA vendors is a VERY bad idea. There are no secrets in the semiconductor ecosystem.
3. See above and once you are on your knees with customers it is VERY hard to get back up.
The most important selling point for a foundry, especially a leading edge foundry, is trust. That should be the first thing IFS thinks about everyday, how do I win the trust of a foundry customer. If you don't have trust you will not succeed in the foundry business. Just ask Samsung.
What is required are tariffs and policies. Tariffs are economy incentives. On top of that, US policy makers need to communicate to fabless companies directly and ask them to use local manufacturing.
Also, trust is a long term goal. For Intel, what they need now is to sort out their fabs/finance asap, i.e., making IFS break even (meaningfully increasing the proportion of Intel 3 and 18A), and enter the AI market meaningfully asap.
siliconbruh999
Well-known member
Trust is a issue for Fabless companies but not for Intel customers they should get thier own customers ownboard first ask them for feedback or way to improve as for PDK I am pretty sure TSMC will be unmatched unless something happens like 10nm, I doubt it cause TSMC doesn't take risk to hang your entire process on something crazy as Co interconnect they will play safe as always it's what make them predictable and a Key thing for building trust.1. IFS does make normal PDKs but they are not as good as TSMC nor will they ever be. TSMC has 30+ years experience making PDKs for customers and TSMC has the top semiconductor companies helping them.
2. Bribing EDA vendors is a VERY bad idea. There are no secrets in the semiconductor ecosystem.
3. See above and once you are on your knees with customers it is VERY hard to get back up.
The most important selling point for a foundry, especially a leading edge foundry, is trust. That should be the first thing IFS thinks about everyday, how do I win the trust of a foundry customer. If you don't have trust you will not succeed in the foundry business. Just ask Samsung.
When I started my career in the 1980s it was said that no one lost their job by choosing IBM. Today it is the same with TSMC for the foundry business.
Hence, Intel should avoid direct competition with TSMC. Just let TSMC say they have the trust advantage or whatever. Intel should leverage on other factors and looking after its own interests should be the priority.
Agreed. I did not take the job from IBM, I joined one of their upstart competitors, a much better way to start a career in Silicon Valley. They (IBM) called us junkyard dogs in suits.
As soon as Pat Gelsinger started with the "We will beat TSMC" narrative I knew the battle was lost. I told Intel privately to go after Samsung first since it was a more realistic battle. I also suggested a "speak softly and carry a big stick" strategy.
I think you will find a completely different narrative under Lip-Bu. He is a tireless promoter yet he is also a down to earth man. My guess is that he is close to billionaire status yet you would never know it when you meet him. I chatted with him at a conference and noticed that he had multiple phones buzzing on and off. That is one busy guy!
tooLongInEDA
Moderator
I really don't believe you can position yourself as weak on trust and succeed in foundry (or indeed most semi business segments). And I'm certain that Lip-Bu Tan doesn't see things that way. It's essential - not a nice to have feature.
If Intel's going to do foundry, it needs to seriously engage a very small number of critical customers to start building that reputation. And in segments where it's really competitive. Something Lip-Bu definitely understands and knows how to do.
I'm also looking forward to Lip-Bu treating TSMC with more respect in public than Pat did. The reality is that there's room for TSMC and Intel to both succeed (particularly if Samsung keeps sliding). It's not a zero sum game (which is historically how Intel seems to have viewed competition).
Avoiding direct competition means not overextending one’s capabilities. PG had an ambitious plan for the foundry, attempting to build fabs worldwide without clearly identifying potential customers. He launched the IFS team with thousands of employees without fully understanding the current market requirements. Such decisions led Intel to its current state, which, in my opinion, was unnecessary. His vision was solid, but the way he executed it was reckless. I also found his management approach contradictory—on one hand, Intel needed to compete with TSMC, but on the other hand, he delayed necessary workforce reductions.
Curious whether Intel should focus more on a competitive Product group or focus on Foundry finances? Hard to see how these could be cooperative goals, but both need to be priorities.
Prioritizing Product means outsourcing or negotiating down Foundry pricing... both of which aren't good for Foundry finances. Prioritizing Foundry means not optimizing processes solely for Product designs and pushing Product to standardize to industry tools & workflows. By separating them, whatever benefit there was to having both under one roof has been severely diminished.
IMO, it should bias towards IDM. Gradually develop foundry business.
hist78
Well-known member
Wait a second.
I just checked the calendar, and today is March 13, 2025, the third decade of the 21st century.
Are we seriously thinking that the IDM model, in leading-edge logic semiconductors, will suddenly work at Intel today, after the entire semiconductor industry has transitioned to the fabless/foundry model?
This transition has taken place over the past 40 years for good reasons. Have any of those reasons changed?
I just checked the calendar, and today is March 13, 2025, the third decade of the 21st century.
Are we seriously thinking that the IDM model, in leading-edge logic semiconductors, will suddenly work at Intel today, after the entire semiconductor industry has transitioned to the fabless/foundry model?
This transition has taken place over the past 40 years for good reasons. Have any of those reasons changed?
hist78
Well-known member
Can you please post your opinions directly on Semiwiki instead of constantly directing people to your Twitter/X account? I understand that you want to build more followers on X and gain higher relevance in Google search, but this makes the Semiwiki forum less efficient and less valuable.
I think posting directly on Semiwiki and mentioning Semiwiki posts on your X account is a simple way to show respect for this forum, bloggers, and its founders.
Not about building followers. It doesn't mean much for me. It is about reusing arguments. A lot of topics have been discussed here a lot of times back and forth.
The path for Intel is simple: to drive up valuation. Deciding whether fab or fabless depends on the context. It is not 1 or 0. Clearly under Trump's administration, fab should be a component. Then the question, how to drive it responsibly. IMO, the objective should be making the valuation of fab assets 1-2x of book value without dragging down the valuation for the product side. At the same time, compete with AI meaningfully.
Analysis: Intel's new CEO should merge Intel Foundry with GF to challenge TSMC's reign
M. S. Lin, a former Fab director at TSMC. Credit: Joseph Chen
Lin believes that Lip-Bu Tan's best course of action to revitalize Intel's foundry business is to emulate the successful strategy of AMD CEO Lisa Su. "Tan should consider merging Intel Foundry Services (IFS) with GlobalFoundries," Lin emphasized. "This would allow Intel to focus on designing and developing its own chip products while fostering a strong, US-based foundry to directly compete with TSMC in the US"
According to Lin, this move would be a game-changer for the industry. "The biggest benefit is that the US would have its own homegrown foundry powerhouse to rival TSMC, which is critical for both industry competition and national security," he explained. "This is also the best strategy for the U.S. to rebuild its semiconductor manufacturing capabilities."
Analysis: Intel's new CEO should merge Intel Foundry with GF to challenge TSMC's reign
Intel may follow AMD's Lisa Su by spinning off its foundry business to GlobalFoundries, creating a US-based chipmaking powerhouse, says a former TSMC executive.www.digitimes.com
in my humble opinion, adding GF to IFS is a bad idea.
GF is not a stable ship, or soon will not be. China's massively overbuilt (and continue building) mature node capacity, once satisfies its internal demand, will soon spill over and be forced to seek business from outside of China, with aggressive pricing to take market share globally everywhere they can. IFS itself will be plenty busy trying to find its first batch of customers and build trust, why do they they want to add on the burden of struggling with the incoming pricing/financial crisis that GF will be facing? Just doesn't make sense to me.
According to the Oregonian, Intel employees (in Oregon) are concerned.
www.oregonlive.com
Intel’s new CEO brings ‘immediate credibility’ but ‘hard decisions’ await
The company's stock jumped 15%.
www.oregonlive.com
Story behind the paywall below,According to the Oregonian, Intel employees (in Oregon) are concerned.
Intel’s new CEO brings ‘immediate credibility’ but ‘hard decisions’ await
The company's stock jumped 15%.
Shortly after Intel announced his hiring Wednesday afternoon, new CEO Lip-Bu Tan greeted the chipmaker’s employees on a conference call promising a “new Intel” focused on pleasing customers and shareholders.
But Tan appeared on the video call wearing a shirt bearing the logo of his old company, Cadence Design Systems, rather than Intel’s name. And he warned his new employees — demoralized by last year’s massive cuts in jobs and spending — that Intel has more “hard decisions” ahead of it, according to two people who listened in.
The 30-minute call did not address the big question facing the troubled company: whether to split Intel’s factories from its chip design business. Instead, Tan pledged to retain Intel’s engineering focus as the company seeks a turnaround.
Some Intel employees had hoped for more enthusiasm from Tan and more details about his plans, according to two workers who spoke with their colleagues and monitored company forums afterwards. There is concern that Tan — who reportedly quit Intel’s board last year because he was frustrated with the company’s bureaucracy — plans to pare back further.
Investors, though, were delighted that Intel’s board had chosen a CEO with deep roots in the semiconductor industry and a track record of success. Tan ran Cadence for a dozen years and has personal connections with former Intel executives and Wall Street analysts.
“Lip-Bu brings immediate credibility to Intel for customers, partners, and investors. He is an icon of the industry with experience spanning decades across both public and private companies,” Bernstein analyst Stacy Rasgon wrote in a note to clients.
Intel’s shares, which had fallen by more than half in the past year, opened 15% higher Thursday morning at $23.80.
“We believe (Intel) has a greater opportunity to restructure/turn things around under his leadership,” Bank of America’s Vivek Arya told clients in a note Thursday. Arya raised his rating on Intel’s stock to “neutral” from “underperform,” saying Tan brings deep expertise and connections throughout the interconnected semiconductor marketplace.
Intel has been on its heels for several years and its sales have fallen by a third since 2021. Tan must overcome years of manufacturing stumbles and rivals who are stealing market share and creating entirely new markets for artificial intelligence.
Former CEO Pat Gelsinger quit suddenly in December, apparently at loggerheads with Intel’s board over the company’s direction. Investment analysts and some former board members have promoted the idea of breaking up the chipmaker, one of the tech industry’s most storied businesses. The financial press has been aflame for months with speculation about potential suitors who might want some of Intel’s pieces.
For the moment, at least, Tan apparently intends to keep Intel intact. He spoke to Intel employees Wednesday about his professional relationships with former Intel executives like Sean Maloney and Dadi Perlmutter, according to people familiar with the call.
“I believe we have a truly unique opportunity to remake our company at one of the most pivotal moments in its history,” Tan told employees in a message posted on Intel’s website.
While Tan spent two years on Intel’s board after leaving Cadence in 2021, he’s the first CEO in Intel’s 57 years who is a true outsider, never having worked at Intel. He quit the company’s board in August, a decision Reuters attributed to his unhappiness with Intel’s bureaucracy and reluctance to cut jobs and costs.
Shortly before his exit, though, Intel announced plans to eliminate 15,000 jobs across the business as the company worked to slash the company’s expenses by $10 billion in 2025. The chipmaker cut 3,000 Oregon jobs through layoffs, buyouts and leaving positions unfilled when employees quit.
Intel remains Oregon’s largest corporate employer, with 20,000 workers at its Washington County campuses.
Now, Tan must stem the bleeding and show that Intel can stand alone — or that there’s a path for its manufacturing and design businesses to thrive independently. Rasgon, the Bernstein analyst, said Tan’s experience in the chip industry and prior position on Intel’s board make him uniquely well positioned to make the difficult choices.
“We view Lip-Bu’s ascension here as positive for Intel, or at least a reason to hope,” Rasgon wrote late Wednesday. “He has a big job in front of him and a lot of wood to chop, though it does seem increasingly likely something will have to change giving investors something more concrete to play for.”
-- Mike Rogoway covers Oregon technology and the state economy. Reach him at mrogoway@oregonian.com or 503-294-7699.