Intel’s CEO: ‘We are not in the top 10’ of leading chip companies

[osnium]

“Twenty, 30 years ago, we are really the leader,” Tan said during a conversation broadcast to Intel employees around the world. “Now I think the world has changed. We are not in the top 10 semiconductor companies.”

Intel’s turnaround will be a “marathon,” Tan said, describing the layoffs that began this week as an effort to make Intel more like rivals Nvidia, Broadcom and AMD, which Tan considers faster and more agile.

“We have to be humble,” Tan said, exhorting staff to listen to customers and respond to their needs. The Oregonian/OregonLive reviewed a recording of his comments.

“There’s a lot of work to do,” Tan said, noting Intel’s loss of market share in data centers. He said its PC business is “doing a bit better,” but he said Intel needs to strengthen its architecture to meet the demands of advanced computing.

“On training I think it is too late for us,” Tan said. He said Nvidia’s position in that market is simply “too strong.”

“Our number one priority is to make sure that our 18A is robust for our internal customer,” Intel’s own processors, Tan said this week. “And then second priority is starting to look at another, 14A, and that’s the next frontier.”

Intel’s CEO: ‘We are not in the top 10’ of leading chip companies

"I think is too late for us" to catch up with Nvidia's position in AI, said Lip-Bu Tan. But he said Intel has other opportunities in the field.

www.oregonlive.com

 

[hist78]

“Twenty, 30 years ago, we are really the leader,” Tan said during a conversation broadcast to Intel employees around the world. “Now I think the world has changed. We are not in the top 10 semiconductor companies.”

Intel’s turnaround will be a “marathon,” Tan said, describing the layoffs that began this week as an effort to make Intel more like rivals Nvidia, Broadcom and AMD, which Tan considers faster and more agile.

“We have to be humble,” Tan said, exhorting staff to listen to customers and respond to their needs. The Oregonian/OregonLive reviewed a recording of his comments.

“There’s a lot of work to do,” Tan said, noting Intel’s loss of market share in data centers. He said its PC business is “doing a bit better,” but he said Intel needs to strengthen its architecture to meet the demands of advanced computing.

“On training I think it is too late for us,” Tan said. He said Nvidia’s position in that market is simply “too strong.”

“Our number one priority is to make sure that our 18A is robust for our internal customer,” Intel’s own processors, Tan said this week. “And then second priority is starting to look at another, 14A, and that’s the next frontier.”

Intel’s CEO: ‘We are not in the top 10’ of leading chip companies

"I think is too late for us" to catch up with Nvidia's position in AI, said Lip-Bu Tan. But he said Intel has other opportunities in the field.

www.oregonlive.com

It's a brutal but honest assessment that serves as a wake-up call for everyone.

 

[Daniel Nenni]

It's a brutal but honest assessment that serves as a wake-up call for everyone.

He is not wrong. It does sound like the Intel Foundry business will skip 18A and run with 14A. The question is will it be with HNA EUV or without? I'm guessing without, the low cost version, go get that NOT TSMC business!

 

[siliconbruh999]

Well they are firing people from their core teams how I am having the same feeling as the reason Sapphire Rapids got delayed laying off people from core team can wait or you just create problems.

https://www.reddit.com/r/intel/s/pB8tdMK52L

 

[XYang2023]

Well they are firing people from their core teams how I am having the same feeling as the reason Sapphire Rapids got delayed laying off people from core team can wait or you just create problems.

https://www.reddit.com/r/intel/s/pB8tdMK52L

View attachment 3346

That could also mean Diamond Rapids is more or less done. Hence, the pull their resources for the next one.

 

[siliconbruh999]

That could also mean Diamond Rapids is more or less done. Hence, the pull their resources for the next one.

Intel's history say otherwise why do you think Clearwater Forest suffered a delay I can bet part of the reason was due to the layoffs last year and if it was finished it was launching this year.
Also Sapphire Rapids where they fired the validation team and it sea of bugs

 

[dkr1986]

Well they are firing people from their core teams how I am having the same feeling as the reason Sapphire Rapids got delayed laying off people from core team can wait or you just create problems.

I have been seeing this on LinkedIn too. 20+yrs experienced engineers leaving Intel saying their position is now eliminated. Layoffs are always messy no matter how targeted a management claim they are going to do it. The company usually ends up rehiring bunch of people they laid off initially as they realize they messed up. It is always disruptive. I feel like if they repeat a Sapphire Rapids again, they are done for in DC! All one can do now is hope the dust settles before the launch window for these 18A products arrive.

Intel's history say otherwise why do you think Clearwater Forest suffered a delay I can bet part of the reason was due to the layoffs last year and if it was finished it was launching this year.
Also Sapphire Rapids where they fired the validation team and it sea of bugs

I think Arrow Lake launch was affected by the 2024 layoffs as well. I feel like CWF on the other hand is different. It is the first time they are doing Hybrid Bonding/ stacked cache in scale and probably ran into some issues which aligns with management commentary about packaging issues.

 

[hist78]

He is not wrong. It does sound like the Intel Foundry business will skip 18A and run with 14A. The question is will it be with HNA EUV or without? I'm guessing without, the low cost version, go get that NOT TSMC business!

I believe Mr. Tan is more concerned about the Intel Product division than Intel Foundry, because the majority of Intel’s actual revenue and profit still come from its products. Right now, Intel’s products are under attack from all directions, and in some cases, the company doesn’t even have a meaningful offering to compete.

One of Pat Gelsinger’s major missteps was devoting so much attention, resources, and billions in capital to Intel’s manufacturing side, while failing to improve the competitiveness and strength of the Product division. That has hurt Intel more than anything else.

 

[dkr1986]

One of Pat Gelsinger’s major missteps was devoting so much attention, resources, and billions in capital to Intel’s manufacturing side, while failing to improve the competitiveness and strength of the Product division.

I don't think this is true. He inherited a very uncompetitive product road map from Swan like Saphire Rapids & Meteor Lake. And also, all the products that were in the drawing board during his tenure are all coming out next couple of years including Panther Lake, Wild Cat Lake, Nova Lake, Diamond Rapids and Clear Water Forrest. Reception of these products will determine whether Pat helped products team or not.

On Client Side (as of now) -
Lunar Lake which was pulled in at least a quarter or two and increased from a niche low volume to major product to compete with Qualcomm's Snapdragon X Elite CPUs. It is a best product for thin and light segment widely recommended by reviewers across the board.
Arrow Lake-S desktop for gaming is a letdown. But it will sell well for OEM desktops. Even for gaming it is decent enough. Only thing here is the lack of x3D competitor.
Arrow Lake-H notebook chips are very competitive with Zen 5 based notebooks including on battery life. Again, lot of recommendations from reviewers.
Battlemage dGPUs received very well.

On Server Side (as of now)-
Emerald Rapids is much better than Sapphire rapids. Granite rapids is even better & closed the gap a ton with AMD offering and is selling well (as well as it can be in an AI GPU dominated DC market). I don't think Sierra Forrest achieved much.

Under Pat Gelsinger, Intel has become more competitive on its core offerings. Only thing you can pin him on is failing to capitalize AI which proved to be devastating imo.

 

[hist78]

I don't think this is true. He inherited a very uncompetitive product road map from Swan like Saphire Rapids & Meteor Lake. And also, all the products that were in the drawing board during his tenure are all coming out next couple of years including Panther Lake, Wild Cat Lake, Nova Lake, Diamond Rapids and Clear Water Forrest. Reception of these products will determine whether Pat helped products team or not.

On Client Side (as of now) -
Lunar Lake which was pulled in at least a quarter or two and increased from a niche low volume to major product to compete with Qualcomm's Snapdragon X Elite CPUs. It is a best product for thin and light segment widely recommended by reviewers across the board.
Arrow Lake-S desktop for gaming is a letdown. But it will sell well for OEM desktops. Even for gaming it is decent enough. Only thing here is the lack of x3D competitor.
Arrow Lake-H notebook chips are very competitive with Zen 5 based notebooks including on battery life. Again, lot of recommendations from reviewers.
Battlemage dGPUs received very well.

On Server Side (as of now)-
Emerald Rapids is much better than Sapphire rapids. Granite rapids is even better & closed the gap a ton with AMD offering and is selling well (as well as it can be in an AI GPU dominated DC market). I don't think Sierra Forrest achieved much.

Under Pat Gelsinger, Intel has become more competitive on its core offerings. Only thing you can pin him on is failing to capitalize AI which proved to be devastating imo.

Yes, Pat Gelsinger inherited many problems when he became CEO of Intel in 2021.

But after 3 years and 10 months in the role before being fired, with shrinking revenue and missed opportunities, how much longer can we say that Gelsinger was not yet fully responsible for Intel’s revenue, profit, and loss?

 

[osnium]

Yes, Pat Gelsinger inherited many problems when he became CEO of Intel in 2021.

But after 3 years and 10 months in the role before being fired, with shrinking revenue and missed opportunities, how much longer can we say that Gelsinger was not yet fully responsible for Intel’s revenue, profit, and loss?

I would argue that some of the spending that Gelsinger pursued for manufacturing did not actually help the fab. Spending money on shells in Germany and Israel for example.

 

[Daniel Nenni]

I would argue that some of the spending that Gelsinger pursued for manufacturing did not actually help the fab. Spending money on shells in Germany and Israel for example.

How much money did Pat G spend on HNA-EUV? Just so he could brag about being first?

 

[dkr1986]

Yes, Pat Gelsinger inherited many problems when he became CEO of Intel in 2021.

But after 3 years and 10 months in the role, with shrinking revenue and missed opportunities, how much longer can we say that Gelsinger is not yet fully responsible for Intel’s revenue, profit, and loss?

I am not saying he is not the one to blame but just there are nuances here.

Client Business - Intel's 2020 & 2021 revenue are result of COVID time govt handouts/ WFH PC purchasing rush. After that it's a cyclical downturn due to PC market slump. This is the case for the last 3 years and affected AMD too. More so for Intel because they are dominant player and more enterprise focused. If you are an IDM with wafer volume dropping, you are in a bigger trouble than a fabless. This is exactly what transpired. Cyclical business being cyclical imo. I mean if you want to blame him for not forecasting that industry trend, over-hiring during that time and being slow to cut head-count. I am with you on that.

Data Center - First blow to Intel's Datac Center CPU business came in when the hyperscalers extended their useful life of servers during COVID time. This led to demand slow down. Second blow is rise of ChatGPT in end of 2022. IMO, Intel had nothing to capitalize on this. All the hysperscalers investment moved to GPUs and more competition from AMD for the remaining pie in cloud. Intel had the in famous SPR to compete during that time, EMR followed on and now GNR a more competitive product. But most of the revenue loss here is due to Saphire Rapids not being competitive vs AMD's Genoa and emergence of AI GPU demand rather than anything Pat Gelsinger did or not do. They lost about $5+Billion in annual revenue on this alone.

Intel had only the MAX GPUs (PVC) and Gaudi lineup to compete. Intel chose Gaudi to compete and cancelled the PVC successor Rialto Bridge, and it was not up to the mark. I am not an expert on this, but I hear more knowledgeable people on this saying they should have chosen the Rialto Bridge instead. I think Intel thought Gaudi is more time to market friendly vs Rialto Bridge (just guessing here), but it turned out to be less competitive vs GPGPUs. If you want to blame that on Pat Gelsinger, I am with you on that partially even though I think Intel never had a chance in AI and still I think that will be the case. Intel's core competency is CPU, only now they are developing their GPU stuff.

Fab/Foundry - The fact some of Intel's own products have moved to TSMC alone is an increase in cost to Intel affecting profit of the company. Fabs are fixed cost and underutilization destroys profit margins. These are decisions made before Pat even became a CEO. But this decision to outsource was a good one as Intel needed to catchup on node processes and this was the only stop gap solution. This outsourcing + accelerated investment increased the expenditures for foundry and became a drag on P&L for Intel to this day. I don't think you can blame him for any of this. Either you go fabless, or you go through this catchup period to be IDM 2.0, no other way.

External Foundry expansion - This is where very valid criticism can be laid out against Pat. He was overambitious and overspent on building fab capacity (e.g. Israel). He was looking to expand further into Germany \ Malayasia but he realized the mistake late & pared back the expansion (too slow to do that as well). But he got CHIPS money for Intel which amounts to $7.86B and another $25B in tax refunds for capex since 2022 (this has increased to >$30B under Trump's big, beautiful bill). He said he underestimated the EDA and IP ecosystem needed to be an external foundry. If you want to blame him for that, I am with you on that.

But none of the above has anything to do with Intel Products competitiveness which I have said improved a ton under him in my previous response.

 

[dkr1986]

I would argue that some of the spending that Gelsinger pursued for manufacturing did not actually help the fab. Spending money on shells in Germany and Israel for example.

There is no shell in Germany, only the land was identified, I don't think they even broke ground iirc. Fab 38 in Israel was supposed to be EUV fab but it was put on hold in 2023 (may be a completed shell now). I think Fab 28 in Israel cannot accommodate EUV machines.

 

[Daniel Nenni]

There is no shell in Germany, only the land was identified, I don't think they even broke ground iirc. Fab 38 in Israel was supposed to be EUV fab but it was put on hold in 2023 (may be a completed shell now). I think Fab 28 in Israel cannot accommodate EUV machines.

I wonder what Lip-Bu will do with the Ohio fabs?

 

[hist78]

I wonder what Lip-Bu will do with the Ohio fabs?

I believe Mr. Li-Bu Tan may allow the Intel Ohio project to remain idle for as long as possible, keeping expenditures minimal. Given Intel’s current lack of near-term demand for the Ohio fab and its constrained financial position, there is little strategic urgency to accelerate construction. However, if Intel decides to split its business, separating Intel Foundry from its product divisions, the Ohio fab may take on a new strategic value and potentially see renewed effort to bring it to live.

 

[MKWVentures]

Pat focused the company on foundry. that was all Pat..... the plan before Pat was to use foundry, not be one. Foundry had 4 years of prep work and cheerleading. No products or tapeouts. People have had many chances and a long time to choose Intel as a foundry. They chose not to.

While Pat was focused on foundry, AI took off. Intel had multiple AI products and programs in the works. Nvidia did incredibly well. AMD is now a 4B/year AI company. Intel is not a top 3 AI company. Intel stock dropped over 50% while others doubled or more. Pat was focused on foundry.

Intel sold off rights or assets on its two leading fabs in order to finance more fabs that they do not need. There are painful clauses in those contracts.

Losing DC market share in CPUs??? I was told Intel was falling behind before Pat. That is the engineers and architects fault. But it didnt get any better under Pat.

Intel can recover by focusing on what they are good at and dropping what they are not good at.... survive the storm, and aligning to be a company at the right size

Pat wanted to be a cheerleader to revive the 1990s (I loved the 90s at Intel). We cannot go back in time. LBT/Intel needs to rebuild for 2025.

Like I said before, the 18A roadmap changes are coming.

 

[count]

A concerning thing for Intel is by many measures they are no longer even as big as thier suppliers and they are much smaller than their customers. This is a company that used to love throwing its weight around, they could push back on suppliers, twist arms if customers and they weren’t shy about it. Even if Intel somehow turns around they will never be in that position of market power again.

 

[swka]

A concerning thing for Intel is by many measures they are no longer even as big as thier suppliers and they are much smaller than their customers. This is a company that used to love throwing its weight around, they could push back on suppliers, twist arms if customers and they weren’t shy about it. Even if Intel somehow turns around they will never be in that position of market power again.

the new CEO, fortunately, does not have that history and culture change baggage. Both external reality and internal top down mandate will hopefully remake culture faster

 

[Barnsley]