Microsoft and Nvidia are now the world’s most valuable companies. How Intel—once the top chipmaker—blew its shot to be one of them

[MKWVentures]

One possible outcome (we have details on multiple possibilities with milestones to track)
Fabless or fab lite. Agreements with people for capacity in exchange for IP (TSMC or GF)
Focus on Architecture leadership. CPUs, GPUs, xPUs, networking. Partnerships with hyperscale companies
Strategic Acquisitions where Intel add value to them
Revenue in the 60B range. Net Profits in the 20B range
smaller size allows Intel to react more quickly and use partners to help with capacity.
Foundry is spun out and renamed to "US Foundry" and government subsidized indefinitely. Some is sold off to GF and TSMC.

Just one of the possible scenarios

 

[osnium]

One possible outcome (we have details on multiple possibilities with milestones to track)
Fabless or fab lite. Agreements with people for capacity in exchange for IP (TSMC or GF)
Focus on Architecture leadership. CPUs, GPUs, xPUs, networking. Partnerships with hyperscale companies
Strategic Acquisitions where Intel add value to them
Revenue in the 60B range. Net Profits in the 20B range
smaller size allows Intel to react more quickly and use partners to help with capacity.
Foundry is spun out and renamed to "US Foundry" and government subsidized indefinitely. Some is sold off to GF and TSMC.

Just one of the possible scenarios

Thank you for the details. When it comes to US Foundry, can you just foresee Intel “selling” the fabs to GF for negative money similar to what IBM did and then having a bigger US foundry as a result? Can you also foresee the government giving up and shutting down “US foundry”? Both scenarios are grim for existing Intel Foundry employees from an employment perspective.

 

[delong.height]

Thank you for the details. When it comes to US Foundry, can you just foresee Intel “selling” the fabs to GF for negative money similar to what IBM did and then having a bigger US foundry as a result? Can you also foresee the government giving up and shutting down “US foundry”? Both scenarios are grim for existing Intel Foundry employees from an employment perspective.

Why would Intel sell fabs to GF when the fabs' book value in Ohio may exceed GF's market cap?

 

[delong.height]

One possible outcome (we have details on multiple possibilities with milestones to track)
Fabless or fab lite. Agreements with people for capacity in exchange for IP (TSMC or GF)
Focus on Architecture leadership. CPUs, GPUs, xPUs, networking. Partnerships with hyperscale companies
Strategic Acquisitions where Intel add value to them
Revenue in the 60B range. Net Profits in the 20B range
smaller size allows Intel to react more quickly and use partners to help with capacity.
Foundry is spun out and renamed to "US Foundry" and government subsidized indefinitely. Some is sold off to GF and TSMC.

Just one of the possible scenarios

It would've been a huge disappointment to me both personally and as an investor owns a significant stake in Intel believe in the story lays in product, then foundry. 60B sounds incredibly low number since Intel has been able to achieve that revenue number since Sept 2017 until March 2023, and that's not without reasons for it to drop. PC weakness, product uncompetitive, process node uncompetitiveness, and now data center weakness due to AI accelerators.

My base line for Intel at 2028 is to be at least 80B, and depending on political environment, 95B and upward. The process node and product uncompetitiveness will be gone by the third quarter of this year. This leaves data center weakness on the table, which eventually it will be normalized. Net profits should be significantly higher even if we assume the revenue is going to be at its highest its ever been ( which is 78.2B in 2021, with net profit at close to 20B) The efficiency policy we've seen on the foundry and design side will save up tons of the cost, such as reducing steppings, reduce hot lots. Being transparent and match the best of the industry. I think it's 25B at least at 78 B and >30B if assume a 90B revenue runway a year.

 

[kingmouf]

It would've been a huge disappointment to me both personally and as an investor owns a significant stake in Intel believe in the story lays in product, then foundry. 60B sounds incredibly low number since Intel has been able to achieve that revenue number since Sept 2017 until March 2023, and that's not without reasons for it to drop. PC weakness, product uncompetitive, process node uncompetitiveness, and now data center weakness due to AI accelerators.

My base line for Intel at 2028 is to be at least 80B, and depending on political environment, 95B and upward. The process node and product uncompetitiveness will be gone by the third quarter of this year. This leaves data center weakness on the table, which eventually it will be normalized. Net profits should be significantly higher even if we assume the revenue is going to be at its highest its ever been ( which is 78.2B in 2021, with net profit at close to 20B) The efficiency policy we've seen on the foundry and design side will save up tons of the cost, such as reducing steppings, reduce hot lots. Being transparent and match the best of the industry. I think it's 25B at least at 78 B and >30B if assume a 90B revenue runway a year.

I do not really know. I think Intel is overselling the new fabs in the context that it is trying to get as many billions as it can get from anyone (me included) that is willing to say "maybe we can build a fab here". But I do not see the contracts supporting all this effort, I do not see any significant clients queueing for whom to build such large capacity and I can see one project being delayed after the other. Intel is adding external investors in the fabs and this means that profits will be shared and it is possible that Intel may not have the freedom to do whatever it may want with the fabs. This is a fundamental shift. As is also a fundamental shift in Intel business to spend billions and billions to create the infrastructure to have low margins compared to what it had in the past. Furthermore, compared even to a few years before, the competitive landscape is much harder now. ARM is everywhere but the problem is not exactly that.... the problem is that the entities that are the big Intel clients are now all working with ARM and producing custom solutions. So it is not that Intel is just loosing ground from a competitor that may (or may not) be producing competitive designs (e.g. AMD), it is actually loosing sales because it is competing with its major customers for all the products that have high margins.
Another thing that I consider dangerous for Intel's future is the fact that it may also loose the China market and anything in the BRICS or now third (and soon possibly second or first) world markets. US government may be willing to give billions to Intel to build US infrastructure but loosing meaningful access to markets for billions of people is not something that can really compensate for that. One can argue that this danger is valid for all Intel's competitors but at this point we are talking about Intel getting from the $60B to a $90B mark in just a few years.
So, while Intel may be able to turn the ship, I would be very cautious to be as optimistic. I believe that it is almost certain that we will see Intel splitting up the company - Intel Design and Intel Foundry (or US foundry as somebody called it in this thread).

 

[BruceA]

Then need to deliver. The world is hungry for advanced technology capacity. They need to fix their PDK quality, culture and delivery to their schedule and the customers will be there. It hard just delivery, but we can see Samsung has struggled and not for lack of money or resources so let’s see….

 

[hist78]

One possible outcome (we have details on multiple possibilities with milestones to track)
Fabless or fab lite. Agreements with people for capacity in exchange for IP (TSMC or GF)
Focus on Architecture leadership. CPUs, GPUs, xPUs, networking. Partnerships with hyperscale companies
Strategic Acquisitions where Intel add value to them
Revenue in the 60B range. Net Profits in the 20B range
smaller size allows Intel to react more quickly and use partners to help with capacity.
Foundry is spun out and renamed to "US Foundry" and government subsidized indefinitely. Some is sold off to GF and TSMC.

Just one of the possible scenarios

"Revenue in the 60B range. Net Profits in the 20B range".

I think it's a tall order for Intel to achieve by 2028. Looking back to Intel net profit margin since 2006, it never went above 30% and that's during a period of time when Intel was the dominant player in the semiconductor market. Now the industry has changed a lot. On the AI/Server processors, there are AMD, Nvidia, Google, Microsoft, Amazon, Apple, Facebook, Tesla, Ampere Computing, Cerebras, Arm, Broadcom,and Qualcomm competing against each other, including Intel. On the client/mobile/embedded/automotive/industrial/communication market, there are Apple, AMD, Nvidia, Broadcom, Tesla, Qualcomm, TI, MediaTek, Amazon, Marvell, etc., in addition to Intel. Many Intel's competitors are much bigger, stronger, more innovative, and good in execution than Intel.

There's a chance for Intel's net profit to go above 30% but I believe that means Intel needs to shrink itself further and go fabless.

Intel Net Profit Margin:

 

[MKWVentures]

"Revenue in the 60B range. Net Profits in the 20B range".

I think it's a tall order for Intel to achieve by 2028. Looking back to Intel net profit margin since 2006, it never went above 30% and that's during a period of time when Intel was the dominant player in the semiconductor market. Now the industry has changed a lot. On the AI/Server processors, there are AMD, Nvidia, Google, Microsoft, Amazon, Apple, Facebook, Tesla, Ampere Computing, Cerebras, Arm, Broadcom,and Qualcomm competing against each other, including Intel. On the client/mobile/embedded/automotive/industrial/communication market, there are Apple, AMD, Nvidia, Broadcom, Tesla, Qualcomm, TI, MediaTek, Amazon, Marvell, etc., in addition to Intel. Many Intel's competitors are much bigger, stronger, more innovative, and good in execution than Intel.

There's a chance for Intel's net profit to go above 30% but I believe that means Intel needs to shrink itself further and go fabless.

Intel Net Profit Margin:

View attachment 2034

My scenario is based on a lean Intel focused on compute with minimal manufacturing ( It would be way different than Intel today). just a scenario for consideration