What's Your Investment Goal with Intel?

[Fabian.]

As an Intel shareholder, I'm curious to know what goals others have when investing in this company.

Let's face it: Intel hasn't been a stellar performer over the past few decades. If you had invested in 1997, you'd likely find yourself in a similar position today, with the stock largely moving sideways.

Despite this, I believe Intel is currently undervalued and has the potential to deliver stronger products in the coming years. My primary reason for investing in Intel, however, is tied to the success of their Foundry 2.0 initiative. I see this as a crucial response to the TSMC monopoly and the looming Taiwan crisis, with China's ambitions regarding Taiwan being well-known. The U.S. CHIPS Act reflects these geopolitical concerns.

Recently, rumors have emerged that Intel Foundry Services (IFS) is consuming so much capital that their ambitious global foundry expansion might slow down or, in extreme cases, be canceled altogether.

Wallstreet liked this and the Stock went 9% up. I, on the other hand, am disappointed about this possible development.

Intel has been in talks with potential foundry customers for some time now and has already lost a potential customer (Broadcom) due to insufficient production capacity and speed.

If they fail to build all the planned leading-edge fabs-like the one in Magdeburg they may not get enough future customers.

Without the full execution of Foundry 2.0, my interest in Intel diminishes significantly. While the risk of bankruptcy or losing market share might decrease, the future growth potential would also be much lower, making other semiconductor stocks with better growth prospects more attractive.

So, what's your investment goal with Intel, especially considering the stock's performance over the past 20 years?

 

[delong.height]

My goal is to get my money back. Then I will see what I'm gonna do base on the company's execution. But so far, not good. And We all know

 

[kevin01]

I am with you on the judgement that IFS has the biggest upside among Intel's various businesses. I remain hopeful that Intel can turn itself around.

Now, my opinion may be not popular: Pat Gelsinger's plan largely worked, with two things he did not anticipate or plan well:

1. Intel's business experienced a meaningful slowdown since the end of Covid-19. Then enterprise/hyperscalers's strong focus on AI investments also worked against Intel.

2. As Pat admitted recently, Intel did not prepare well enough on EDA and IP tools for potential fab customers.

Except for these two things, they seem to have largely delivered 5N4Y -- a significant technical achivement that few people expected them to accompanish, especially after their poor track record in the 2010s decade.

Going forward, here is a possible bull case for Intel:

1. Intel's business in CCG and DC servers have some decent growth for a few years (compared to where they are in 2024), due to strong product offerings and/or change of business cycle.

2. Intel can finish at least one EUV fab in the U.S. in 2025 and 2026. currently they only have one production fab in Ireland.

3. Intel achieves reasonally good yield on 18A.

4. Intel can have a small technical lead in future foundry offerings after 18A.

I think that 1 and 2 should happen with high probability; 3 and 4 are not certain, but it is a fair game against TSMC.

Many investors are concerned with the seemingly lack of the foundry customers. But the current situation is very understandable, because IFS simply does not have the track record yet. Once it becomes clear that 18A is a winner, IFS will have a much easier time to find external customers. Now all they need to do is to deliver 18A, and Intel products can fill its current capacity -- in fact, even in Intel's plan, their goal is, by 2030 (or 2027?), to have 2/3 of IFS manufacture for Intel products and 1/3 for external customers.

 

[delong.height]

I am with you on the judgement that IFS has the biggest upside among Intel's various businesses. I remain hopeful that Intel can turn itself around.

Now, my opinion may be not popular: Pat Gelsinger's plan largely worked, with two things he did not anticipate or plan well:

1. Intel's business experienced a meaningful slowdown since the end of Covid-19. Then enterprise/hyperscalers's strong focus on AI investments also worked against Intel.

2. As Pat admitted recently, Intel did not prepare well enough on EDA and IP tools for potential fab customers.

Except for these two things, they seem to have largely delivered 5N4Y -- a significant technical achivement that few people expected them to accompanish, especially after their poor track record in the 2010s decade.

Going forward, here is a possible bull case for Intel:

1. Intel's business in CCG and DC servers have some decent growth for a few years (compared to where they are in 2024), due to strong product offerings and/or change of business cycle.

2. Intel can finish at least one EUV fab in the U.S. in 2025 and 2026. currently they only have one production fab in Ireland.

3. Intel achieves reasonally good yield on 18A.

4. Intel can have a small technical lead in future foundry offerings after 18A.

I think that 1 and 2 should happen with high probability; 3 and 4 are not certain, but it is a fair game against TSMC.

Many investors are concerned with the seemingly lack of the foundry customers. But the current situation is very understandable, because IFS simply does not have the track record yet. Once it becomes clear that 18A is a winner, IFS will have a much easier time to find external customers. Now all they need to do is to deliver 18A, and Intel products can fill its current capacity -- in fact, even in Intel's plan, their goal is, by 2030 (or 2027?), to have 2/3 of IFS manufacture for Intel products and 1/3 for external customers.

But one thing that recently discourage me the most, is the fact that they may not be able to be the second in revenue if they failed to capture sufficient volume because they don't have much supply. I think the demand for leading-edge process node still outweight the supply of them. But Intel cannot afford to build one more fab, so TSMC may still capture 85% of the market. Whereas Intel and Samsung share the remaining 15%. The pace of expansion has dramatically slowed down compare to the assumption I first made a couple years ago.

 

[VCT]

Intel's design worth $40 per share alone but they need to spin off manufacturing.
I plan to buy in 2025-2026 at a cost of $15-$19 and wait for spin off.

 

[delong.height]

Intel's design worth $40 per share alone.
I plan to buy in 2025-2026 at a cost of $15-$19 and wait for spin off.

I think it worth $60 per share. Lunar lake has demonstrated that they are worthy of fighting Qualcomm on certain low power devices, they can go after AMD on console, go after Qualcomm on car and AR/VR headset. They definitely need more experience building low-power but powerful chip. Intel product should be worth at 300B assuming low CAGR and healthy yet competitive product portfolio which they should be able to get to by 2025. It's largely profitable.

 

[VCT]

I think it worth $60 per share. Lunar lake has demonstrated that they are worthy of fighting Qualcomm on certain low power devices, they can go after AMD on console, go after Qualcomm on car and AR/VR headset. They definitely need more experience building low-power but powerful chip.

agree. $40 is easy and $60-80 is highly possible.
If Intel make their production back from TSMC in 2027 that means IFS is good enough to spin off.
If not, Pat is going to step down any way and new CEO will spin off Intel as well.

 

[delong.height]

agree. $40 is easy and $60-80 is highly possible.
If Intel make their production back from TSMC in 2027 that means IFS is good enough to spin off.
If not, Pat is going to step down any way and new CEO will spin off Intel as well.

He shall step down soon by 2025, or ideally, that is the year when he should retire (65). But that man wanted a few more years as CEO back in 2021. And after all these year, these suffering and criticism from wall street and Intel internally, I wonder if he still want to be.

 

[kevin01]

I think it worth $60 per share. Lunar lake has demonstrated that they are worthy of fighting Qualcomm on certain low power devices, they can go after AMD on console, go after Qualcomm on car and AR/VR headset. They definitely need more experience building low-power but powerful chip. Intel product should be worth at 300B assuming low CAGR and healthy yet competitive product portfolio which they should be able to get to by 2025. It's largely profitable.

Intel management admits that lunar lake has lower profit margins, because it uses TSMC leading edge process, which is expensive. After all, amd and nvda chose to use leading-1 nodes for good reasons. Only really high margin products like the newest iphones or the best nvda AI chips can afford the best manufacturing processes.

Do you still think Intel design is worth $60 alone?

 

[delong.height]

Intel management admits that lunar lake has lower profit margins, because it uses TSMC leading edge process, which is expensive. After all, amd and nvda chose to use leading-1 nodes for good reasons. Only really high margin products like the newest iphones or the best nvda AI chips can afford the best manufacturing processes.

Do you still think Intel design is worth $60 alone?

They should be. It is lower profit margin, but margin is a percentage. Lunar lake was sold as a memory attached SoC with either 16GB or 32GB. So even if the profit margin decline from say '50% to 40%". The total gross profit should be increasing as well, but price to OEM is much higher, because it included memory already and it's packaged up, saving real estate of the board and thermal. Intel resell the memory chip at a price higher than they get. The cost of having it fab at TSMC, the cost of acquire memory, and the cost of packaging them up. The price should be much higher, and I believe OEM is happy to take it at face value because it helps them to sell more units against Apple.

Honestly, Qualcomm Snapdragon elite is good, but because of compatibility, it will never let OEM feel comfortable against Apple. Lunar lake is probably the best in its class.

 

[XYang2023]

Intel management admits that lunar lake has lower profit margins, because it uses TSMC leading edge process, which is expensive. After all, amd and nvda chose to use leading-1 nodes for good reasons. Only really high margin products like the newest iphones or the best nvda AI chips can afford the best manufacturing processes.

Do you still think Intel design is worth $60 alone?

I think the valuation is reasonable. But Intel should not shedding away assets like Mobileye and Altera, especial like what PG is doing now, shedding them at depressed values. Taiwanese media described the current rumours with words like "spendthrift", "drug-addicts" (chasing fabs), etc.

They should follow the examples of Nvidia and AMD to integrate those IPs into next-gen products.

 

[delong.height]

I think the valuation is reasonable. But Intel should not shedding away assets like Mobileye and Altera, especial like what PG is doing now, shedding them at depressed values. Taiwanese media described the current rumours with words like "spendthrift", "drug-addicts" (chasing fabs), etc.

Shedding away assets like Mobileye and Altera is what I hated the most. They don't need the money. Investing in US is not practical long term. You can build 1-2 fabs once a decade for some level of resilience. But IFS should be outsource their 18A to Samsung, which is what I proposed. They just cannot afford to spend a lot of capex into foundry because people don't want to pay higher prices unless there is tax tariff on those chip manufactured oversea and IFS have credible reputation.

They should follow the examples of Nvidia and AMD to integrate those IPs into next-gen products.

Yes they should, but Intel is bloated, integrate them into next-gen products are not desirable because Intel design hasn't achieve the level of efficiency AMD/NVDA has. They still need some time to prove it. But so far so good. Another good example is Qualcomm emphaisis in building more branded product instead of licensing its IP, which has higher profit margin but lower revenue

 

[delong.height]

Shedding away assets like Mobileye and Altera is what I hated the most. They don't need the money. Investing in US is not practical long term. You can build 1-2 fabs once a decade for some level of resilience. But IFS should be outsource their 18A to Samsung, which is what I proposed. They just cannot afford to spend a lot of capex into foundry because people don't want to pay higher prices unless there is tax tariff on those chip manufactured oversea and IFS have credible reputation.


Yes they should, but Intel is bloated, integrate them into next-gen products are not desirable because Intel design hasn't achieve the level of efficiency AMD/NVDA has. They still need some time to prove it. But so far so good. Another good example is Qualcomm emphaisis in building more branded product instead of licensing its IP, which has higher profit margin but lower revenue

What they should be doing instead is return the shares to shareholder, they don't need money to build out more fabs. Seek more strategic options (like partnering with Samsung for their capacity), it's in Asia, so it's cheap. Intel is known for its low ROIC to shareholders. They need to think of ways to elevate it, instead of keeping it, or making it lower.

 

[XYang2023]

What they should be doing instead is return the shares to shareholder, they don't need money to build out more fabs. Seek more strategic options (like partnering with Samsung for their capacity), it's in Asia, so it's cheap. Intel is known for its low ROIC to shareholders. They need to think of ways to elevate it, instead of keeping it, or making it lower.

That is why I think the BoD should let PG retired gracefully.

 

[delong.height]

That is why I think the BoD should let PG retired gracefully.

In the past I would disagree, but now I think that's a good option.

But he should be able to get a great compensation package for what he has achieved, with a chair on the board. He has done excellent jobs in 5N4Y and navigated these difficult situation. He should be there given the dynamic of the industry. He can be a great mentor.

 

[tooLongInEDA]

In the past I would disagree, but now I think that's a good option.

But he should be able to get a great compensation package for what he has achieved, with a chair on the board. He has done excellent jobs in 5N4Y and navigated these difficult situation. He should be there given the dynamic of the industry. He can be a great mentor.

He already has a great compensation package. Would have been spectacular if the stock price hadn't dived under his charge.

Come on, we're engineers. We know it's really 3N4Y at best. Intel 4/3 are really one node with a shrink, as are 20A/18A. And some of these will hardly be used. And it may be >4Y. Still an achievement and a vast improvement on the preceding chaos. But way overstated. The real measure is the revenue generated off these nodes and not hitting internal milestones ("marking your own homework"). And how this measures up against what they got from earlier nodes like 22 and 14.

 

[siliconbruh999]

He already has a great compensation package. Would have been spectacular if the stock price hadn't dived under his charge.

Come on, we're engineers. We know it's really 3N4Y at best. Intel 4/3 are really one node with a shrink, as are 20A/18A. And some of these will hardly be used. And it may be >4Y. Still an achievement and a vast improvement on the preceding chaos. But way overstated. The real measure is the revenue generated off these nodes and not hitting internal milestones ("marking your own homework"). And how this measures up against what they got from earlier nodes like 22 and 14.

I don't think anything in history comes close to 14nm like it was ridiculous 14nm+++++

 

[count]

If/when Intel exits the fab business the stock will go back up to maybe $30, and the company can start to become healthy again. The longer they wait the harder it will be for shareholders.

 

[Mooredaddy]

agree. $40 is easy and $60-80 is highly possible.
If Intel make their production back from TSMC in 2027 that means IFS is good enough to spin off.
If not, Pat is going to step down any way and new CEO will spin off Intel as well.

Hopium unfortunately

 

[samwilde]

Shedding away assets like Mobileye and Altera is what I hated the most. They don't need the money.

I would say I agree with this. They will get a limited amount of money by selling these assets under duress. And much like selling off XScale it will kick Intel out of what are important markets. In the case of Mobileye this is self-driving vehicles, and in the case of Altera it could be used as a great way to drive fab volume.

Investing in US is not practical long term. You can build 1-2 fabs once a decade for some level of resilience.

I disagree here. With state support similar to that provided in Asia there is little reason to expect costs in the US to be a significant issue.

But IFS should be outsource their 18A to Samsung, which is what I proposed. They just cannot afford to spend a lot of capex into foundry because people don't want to pay higher prices unless there is tax tariff on those chip manufactured oversea and IFS have credible reputation.

Intel needs like four fabs to produce all its chip demand. I agree with you guys that they already have the Irish fab, and will likely finish one of the US fabs. As for the other two US fabs, or some other place fabs, this will depend on execution and capital availability.

I agree with your plan of licensing 18A to Samsung. Short term it will likely be impossible for Intel to finance those fabs. They could could have financed the four fabs if they had cut dividends two years before. They should have used the bonanza of the lockdown sales to build those fabs.