Exclusive: Intel's new CEO explores big shift in chip manufacturing business

[user nl]

Intel does manufacture advanced 3D packaging.

Unclear until a chip packaging expert posts a deep-dive here or on similar sites.

Huh? The speed of electrical transmission through copper is 95% of the speed of light, so for all practical purposes at the circuit level optics do not transmit signals significantly faster than electricity. Also optics are only more energy efficient with silicon photonics, which are nascent. If you need optical transceivers you need more power, have more cost, and higher latency.

Perhaps follow or read some on this substack site:
https://tspasemiconductor.substack.com/

 

[user nl]

Intel does manufacture advanced 3D packaging.

Unclear until a chip packaging expert posts a deep-dive here or on similar sites.

Huh? The speed of electrical transmission through copper is 95% of the speed of light, so for all practical purposes at the circuit level optics do not transmit signals significantly faster than electricity. Also optics are only more energy efficient with silicon photonics, which are nascent. If you need optical transceivers you need more power, have more cost, and higher latency.

Maybe a good 2025 reference for beginners like me (I'm no semi-expert, just a simple physicist) is (see also references in this 2025 Review):

Optical network-on-chip (ONoC) architectures: a detailed analysis of optical router designs
Yasin Asadi 2025 J. Semicond. 46 031401
https://www.jos.ac.cn/en/article/doi/10.1088/1674-4926/24060006


9. Conclusions
In conclusion, the emergence of ONoC technology offers a promising solution to address the limitations of traditional electronic interconnects, such as bandwidth constraints, high latency, and power consumption issues. Within ONoC design architectures, optical routers play a critical role by enabling high-speed, low-latency data transfer, efficient data routing, and meeting the scalability requirements of modern computing systems. This study serves as an introductory resource for beginners, providing a fundamental understanding of ONoC and optical routers, while also offering a comprehensive survey and analysis for experts. It focuses on three main categories of router architectures: microring resonators (mrrs), MZIs, and hybrid designs. Additionally, the study provides insights into the specific advantages and limitations of these router architectures.

 

[blueone]

Perhaps follow or read some on this substack site:
https://tspasemiconductor.substack.com/

I don't pay much attention to that site. The articles are largely biased marketing.

 

[blueone]

I'm no semi-expert, just a simple physicist

In my experience, there is no such thing as a simple physicist.

 

[Artificer60]

Curious... did LBT get his US citizenship yet?

According to google he got his citizenship in 1987. So yeah, quite a while ago.

"Yes, Lip-Bu Tan is a naturalized U.S. citizen, according to news articles and his biography from Intel. He was born in Malaysia and raised in Singapore before coming to the United States for his education. He became a naturalized citizen in 1987"

 

[Artificer60]

Folks argue that Intel is important to the US given that it’s one of two us based companies doing leading edge logic semiconductor research (the other being IBM research). Given the broken culture, terrible execution and unsustainable finances, should the US government just give up, withdraw CHIPs act money for Intel and beg TSMC to be more aggressive about ramping leading edge nodes faster in Arizona. It seems like Intel Foundry makes Boeing look like a competent and well run manufacturer.

IBM semi research is a joke. Their processes don't yield. There is a world of difference between a research paper and running actual silicon in high volume.

Not sure why you keep advocating for the US to seed the most important industry in the foreseeable future to some other country.

 

[osnium]

IBM semi research is a joke. Their processes don't yield. There is a world of difference between a research paper and running actual silicon in high volume.

Not sure why you keep advocating for the US to seed the most important industry in the foreseeable future to some other country.

Under pure capitalism, if a company is not competitive, it goes under. The concern I have is that Intel’s employees don’t have an incentive to reform themselves and execute better if they hade a sense that they are too big to fail. It creates a moral hazard where they execute poorly and get bailed out. I agree that it’s a critical industry but the right approach might be to bring in external expertise rather than relying on an organization with a poor track record.

What are your thoughts on how to avoid this sense of moral hazard?

 

[Artificer60]

Under pure capitalism, if a company is not competitive, it goes under. The concern I have is that Intel’s employees don’t have an incentive to reform themselves and execute better if they hade a sense that they are too big to fail. It creates a moral hazard where they execute poorly and get bailed out. I agree that it’s a critical industry but the right approach might be to bring in external expertise rather than relying on an organization with a poor track record.

What are your thoughts on how to avoid this sense of moral hazard?

I don't think Intel employees feel any such thing. When the dust settles on this next round of layoffs ~30% of the employee base will have been laid off in the last 9 months. I can't imagine that anyone who will be around by the end of the month feels that something doesn't need to change. No one bailed out those people who were let go.

I also don't feel the issue is with the employees at the bottom of the org structure. I believe they are smart, capable, and frustrated by a system that prevents them from doing their best work. Remember, this is the same company that was first to implement strained-Si, Hi-K metal gate and fin-fet transistors in HVM and will soon add BSPD to that list. The problem is a management structure that feels like they can dictate their will to the market and handcuffs those employees. If that doesn't change, Intel will die and the US will lose it's best shot at being a player in leading edge logic development.

TSMC is not a solution for the US. They will not be doing leading edge development in the US. The skill set to sustain a factory and make incremental improvements to an existing process is different from the skill set required to build a new process innovation like gate all around transistors. If Intel dies that skill set will no longer exist in the US. The catch 22 here is that I don't think something that the national labs will work here either, because you need volume learning and that means building fabs and churning out thousands of wafers a month. Not what the national labs are designed to do.

To be clear, I'm not advocating for a government bailout of Intel. What I do believe the US needs to do is to provide support to Intel similar to the support Taiwan provides TSMC. Those supports include tax breaks for R&D and purchase of equipment and subsidized rates for utilities. None of those thing feel like a government bailout to me.

The issue at Intel seems to be similar to the issues that Cadence was facing when Lip-Bu Tan took over. I truly believe that he is Intel's best shot to fix their structural issues and pull the company out.

 

[hist78]

Under pure capitalism, if a company is not competitive, it goes under. The concern I have is that Intel’s employees don’t have an incentive to reform themselves and execute better if they hade a sense that they are too big to fail. It creates a moral hazard where they execute poorly and get bailed out. I agree that it’s a critical industry but the right approach might be to bring in external expertise rather than relying on an organization with a poor track record.

What are your thoughts on how to avoid this sense of moral hazard?

I don't think Intel employees feel any such thing. When the dust settles on this next round of layoffs ~30% of the employee base will have been laid off in the last 9 months. I can't imagine that anyone who will be around by the end of the month feels that something doesn't need to change. No one bailed out those people who were let go.

I also don't feel the issue is with the employees at the bottom of the org structure. I believe they are smart, capable, and frustrated by a system that prevents them from doing their best work. Remember, this is the same company that was first to implement strained-Si, Hi-K metal gate and fin-fet transistors in HVM and will soon add BSPD to that list. The problem is a management structure that feels like they can dictate their will to the market and handcuffs those employees. If that doesn't change, Intel will die and the US will lose it's best shot at being a player in leading edge logic development.

TSMC is not a solution for the US. They will not be doing leading edge development in the US. The skill set to sustain a factory and make incremental improvements to an existing process is different from the skill set required to build a new process innovation like gate all around transistors. If Intel dies that skill set will no longer exist in the US. The catch 22 here is that I don't think something that the national labs will work here either, because you need volume learning and that means building fabs and churning out thousands of wafers a month. Not what the national labs are designed to do.

To be clear, I'm not advocating for a government bailout of Intel. What I do believe the US needs to do is to provide support to Intel similar to the support Taiwan provides TSMC. Those supports include tax breaks for R&D and purchase of equipment and subsidized rates for utilities. None of those thing feel like a government bailout to me.

The issue at Intel seems to be similar to the issues that Cadence was facing when Lip-Bu Tan took over. I truly believe that he is Intel's best shot to fix their structural issues and pull the company out.

From Trump Administration 1 to the Biden Administration and now to Trump Administration 2, two policy choices have remained consistent and clear:

1. Bring proven winners to the US to establish more semiconductor manufacturing capacity, as long as they are American companies or come from friendly nations.

2. Intel’s manufacturing division (Intel Foundry), not its product division, is considered worth supporting with government assistance. However, taxpayer money cannot solve the many fundamental problems Intel is facing. In a free market system, the government is neither smart enough nor does it have the authority to dictate what’s right or wrong for Intel. Decisions around Intel’s business model, capital allocation, executive appointments, board oversight, technology direction, product development, and pricing choices are entirely Intel’s own responsibility and destiny.

If Intel’s product division fails, it’s not a major issue, since there are already several strong American competitors outperforming Intel. But what if Intel Foundry fails? The answer lies in policy #1 above.

 

[ai268]

That’s fair about IBM but I think the question remains. Would you advise the US government to continue pouring money into Intel via Chips Act while they struggle or should the government reconsider and reallocate money to TSMC instead?

Tariff man still thinks he is invincible and maybe have NOT even, or never, entered "denial" phase yet. There is nothing man-power & money can not solve, so did China and Samsung assume before and now is US/Intel's turn. US loses to a Pacific-Rim small Island. No way, Jose. Right?

The issue at Intel seems to be similar to the issues that Cadence was facing when Lip-Bu Tan took over. I truly believe that he is Intel's best shot to fix their structural issues and pull the company out.

I just do NOT understand why everybody here put so much bet / faith on Lip-Bu Tan. He just had a nuclear engineering master degree, even from MIT. Then a VC and NOT a contrarian at all, like he claimed. Cadence is a paper-machine builder, software company. IC building is solid-state physics & chemistry. It`s just like Architect vs Civil engineer, similar but totally different. Did Intel want to hire CEO with Foundry expertise? Does Cadence CEO experience fit that category?

 

[benb]

Downsizing, bringing in consultants, doing a merger, splitting the business up, staying the course or government intervention are the big options on the table for Intel.

Intervention is the last option for good reason. It seems like we're not quite there yet, they are currently doing downsize+stay the course. Staying the course ("the next node will deliver, just wait") feels like insanity to me though, sometimes.

It's interesting that almost the same thing is happening at Tesla, SpaceX, Intel, and Boeing. If the US were to lose all of those it would truly be a crisis.

 

[ai268]

Downsizing, bringing in consultants, doing a merger, splitting the business up, staying the course or government intervention are the big options on the table for Intel.

Intervention is the last option for good reason. It seems like we're not quite there yet, they are currently doing downsize+stay the course. Staying the course ("the next node will deliver, just wait") feels like insanity to me though, sometimes.

It's interesting that almost the same thing is happening at Tesla, SpaceX, Intel, and Boeing. If the US were to lose all of those it would truly be a crisis.

How can you expect? There are at least two hedge fund great whale in the cabinet. Long, Short Stock, Treasury, bond, FX, real estate with Fed. QE. Easy money, why manufacturing with 3-4% profit margin and polluting the environment? Are you nuts?

 

[KevinK]

Not sure why you keep advocating for the US to seed the most important industry in the foreseeable future to some other country.

That would be "cede". And I think it would wrong to cede our position.

 

[osnium]

From Trump Administration 1 to the Biden Administration and now to Trump Administration 2, two policy choices have remained consistent and clear:

1. Bring proven winners to the US to establish more semiconductor manufacturing capacity, as long as they are American companies or come from friendly nations.

2. Intel’s manufacturing division (Intel Foundry), not its product division, is considered worth supporting with government assistance. However, taxpayer money cannot solve the many fundamental problems Intel is facing. In a free market system, the government is neither smart enough nor does it have the authority to dictate what’s right or wrong for Intel. Decisions around Intel’s business model, capital allocation, executive appointments, board oversight, technology direction, product development, and pricing choices are entirely Intel’s own responsibility and destiny.

If Intel’s product division fails, it’s not a major issue, since there are already several strong American competitors outperforming Intel. But what if Intel Foundry fails? The answer lies in policy #1 above.

Would could policy #1 entail if Intel does not successfully execute on 14A? TSMC Arizona all the way with some commitment to do some process development in Arizona? Could it involve some sort of merger with GlobalFoundries along with a government golden share akin to Nippon Steel and US Steel? Given all this uncertainty, would you expect any and all competent Intel foundry employees to jump ship?

 

[user nl]

I don't pay much attention to that site. The articles are largely biased marketing.

OK, a real paper then for you, published on 5 June 2025, hopefully not outdated yet in this fast-paced semi-world, you can perhaps tell me ? See the conclusions here:

The Review contains many very nice photos of all kinds of CPO chips/boards/hardware, for a non-semi-expert like me very digestible and simple....Here an Intel CPO-chip from 2021:

For further studies read the Review here (a real engineering journal I would think?)":

Proceedings Volume 13648, Fifth Optics Frontier Conference (OFS 2025); 136480G (2025)

https://doi.org/10.1117/12.3069633

 

[Artificer60]

That would be "cede". And I think it would wrong to cede our position.

I stand corrected.

 

[Artificer60]

I just do NOT understand why everybody here put so much bet / faith on Lip-Bu Tan. He just had a nuclear engineering master degree, even from MIT. Then a VC and NOT a contrarian at all, like he claimed. Cadence is a paper-machine builder, software company. IC building is solid-state physics & chemistry. It`s just like Architect vs Civil engineer, similar but totally different. Did Intel want to hire CEO with Foundry expertise? Does Cadence CEO experience fit that category?

I can't speak for anyone else, but it my case it is because the issue at Intel isn't technical talent. Intel has that in spades. What they need is someone to clear out the bureaucracy and let that talent flourish. If you look hard enough you can find some details on what Lip-Bu Tan did at Cadence and it was precisely the things that Intel needs.

Pat Gelsinger's was an Electrical Engineer and I believe his basic strategy was sound. I think his problem was that he didn't realize what they needed to build before they would come. As a result he spent a ton of money on the wrong thing (increased Fab capacity) and missed out on the importance of the customer experience. Lip-Bu Tan is running with the same strategy, but controlling costs and claims to be making changes to how Intel will interact with their customers, though I have yet to see any signs of the later.

At Cadence Lip-Bu Tan met with his customers, actually listened to what they said they wanted, and delivered that to them. That seems to be the same approach that has won TSMC such loyalty. They listen, they deliver the things their customers want and do it on the timeline they promise. Intel's reputation in the industry does not fit that mold, but If Lip-Bu Tan can change that, Intel will win business, because the technical ability is already there.

 

[siliconbruh999]

I can't speak for anyone else, but it my case it is because the issue at Intel isn't technical talent. Intel has that in spades. What they need is someone to clear out the bureaucracy and let that talent flourish. If you look hard enough you can find some details on what Lip-Bu Tan did at Cadence and it was precisely the things that Intel needs.

Pat Gelsinger's was an Electrical Engineer and I believe his basic strategy was sound. I think his problem was that he didn't realize what they needed to build before they would come. As a result he spent a ton of money on the wrong thing (increased Fab capacity) and missed out on the importance of the customer experience. Lip-Bu Tan is running with the same strategy, but controlling costs and claims to be making changes to how Intel will interact with their customers, though I have yet to see any signs of the later.

At Cadence Lip-Bu Tan met with his customers, actually listened to what they said they wanted, and delivered that to them. That seems to be the same approach that has won TSMC such loyalty. They listen, they deliver the things their customers want and do it on the timeline they promise. Intel's reputation in the industry does not fit that mold, but If Lip-Bu Tan can change that, Intel will win business, because the technical ability is already there.

i am sorry to say but they have lost some talent as well all the ex Intel employees have been poached by Apple/Nvidia