Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/threads/the-three-interlocking-problems-at-intel.23548/
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021770
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

The Three Interlocking Problems at Intel

benb

Well-known member
The Three Interlocking Problems at Intel
BenB


Challenges to Sustainability
-Credit rating. Intel credit rating was downgraded to BBB by Fitch in August 2025.
-IDM business model. This reduces the addressable market for Intel Foundry since AMD and Nvidia are direct competitors.
-Separation costs. It would be costly to separate Intel Products from Intel Foundry. This cost is not specified.
-Going concern risk. Intel has liabilities of $32 billion due within a year, and $53 billion in liabilities outstanding.


Outline of a solution
-Intel needs relief of a majority of the liabilities
-Separate Intel Products and Intel Foundry.
-Intel Foundry needs a sufficient customer base

How to achieve the solution
-Intel recently received $8.9 billion from the US Treasury in return for 10% equity stake. That cash helps but is not enough to relieve the liabilities. The liabilities in turn prevent separating Intel Product and Intel Foundry, since creditors are not all secured. And separating Intel Foundry is most likely needed to increase the addressable market for Intel Foundry, increasing the possibility of finding a sufficient customer base.

Breaking the deadlock
-Intel has large value to the USA in war. In a war between China and the US, Korea and Taiwan might remain neutral and be unwilling (or unable) to supply the US (or China). In an AI-driven drone war, Intel production of advanced nodes could be decisive since advanced AIs would defeat simple AIs. That is the goal of limiting China to N-2 advanced node production.
-While Intel would be valuable in a war, this value remains dormant at peace, and there are roadblocks to unlocking the value. These roadblock include:
-No US advanced drone supply chain
-Nvidia and AMD must produce chips in Taiwan to remain on the most advanced nodes. TSMC Arizona is N-1, better than N-2 but precariously close to it.
-SK Hynix AI HBM chips (input to Nvidia AI) is Korea or China-based
-US power grid may not be able to power advanced AI at a war (or any) scale


Solution
-US Treasury would pay off Intel debts
-This would permit separating Intel Foundry
-In return, Intel Foundry would become largely a government enterprise for the production of AI war supplies. That may mean Nvidia, AMD, Intel and Micron products. US government money would purchase these products and require their production on 14A and other future nodes. That in turn would fill Intel (and Micron) fabs.

Failure to execute this plan
-China achieves AI war dominance. This may have already happened, with Deepseek reducing the hardware requirements for the most advanced AIs.
-Intel stops advanced node development because they don’t separate Intel Foundry, limiting the market for Intel Foundry, leading to no 14A external customers.
-US advanced node production becomes N-1 to Taiwan and Korea. TSMC is only willing to provide N-1 fabs on US soil, which is precariously close to N-2 of China. It may effectively be the same as N-1 if China finds workarounds.
 
It is OK to separate foundry from product team, but foundry should keep intel brand and trademark. It was the foundry that propelled intel for so long. Intel design/product team has not been competitive against peers and is deemed to fail without the close tie it had with foundry team.
 
It is OK to separate foundry from product team, but foundry should keep intel brand and trademark. It was the foundry that propelled intel for so long. Intel design/product team has not been competitive against peers and is deemed to fail without the close tie it had with foundry team.
Foundry is not customer facing so the brand is not relevant. It could be TrumpFoundries and it would not matter to foundry customers. On the other hand, there are still people who look for Intel processors for their laptops. Also, I would argue that it’s the fab’s fault for all the problems that Intel has had for the past decade, starting with 10 nm delays.
 
-While Intel would be valuable in a war, this value remains dormant at peace, and there are roadblocks to unlocking the value. These roadblock include:
-No US advanced drone supply chain
-US power grid may not be able to power advanced AI at a war (or any) scale
An excellent write up btw! The credit rating and debt summary is eye opening.

Just two minor comments/thoughts:

- Hegseth has made a few announcements recently that the DoD is working to build up that US supply chain. The USA could end up with a significant drone supply chain in just 3 years (These aren't super complex devices like semiconductors).

- I'm a skeptic on the US power grid, but -- it was supposed to collapse with millions of EVs sold and it hasn't. The grid is also successfully handling AI so far; and the amount of stationary energy capacity in the US is greatly increasing thanks to products like Tesla megapacks or powerwall equivalents (stationary "on-demand" power storage can allow power plants to produce potentially 2X+ energy than they normally do).

In wartime, the Defense Dept would certainly usurp commercial use of AI if needed. That said, we could be a few years away from other technical limitations grinding us to a halt.
 
Foundry is not customer facing so the brand is not relevant. It could be TrumpFoundries and it would not matter to foundry customers. On the other hand, there are still people who look for Intel processors for their laptops. Also, I would argue that it’s the fab’s fault for all the problems that Intel has had for the past decade, starting with 10 nm delays.
Both can be true --

The product team was developing products based on a constant cadence and supply of new node geometries and performance.

However, this choice also limited their competitiveness when foundry failed to delivery.
 
In a war scenario, we need to be able to produce AMD and Nvidia AI products (and Intel too although secondarily) on node N. Node N will be Intel, TINA. I see no alternative other than to separate Intel and slot AMD and Nvidia in Intel fabs. Intel Products will probably sprint away from Intel Foundry which is fine, none of those products are strategic. It's a bad, bad scenario but for $53 billion, with the numbers people throw around for AI, it's doable.

Not doing it soon will lead to Intel crashing in about 12 months. I think their debt will be junk soon. So they're starting to fail as of August 2025.
 
In a war scenario, we need to be able to produce AMD and Nvidia AI products (and Intel too although secondarily) on node N. Node N will be Intel, TINA. I see no alternative other than to separate Intel and slot AMD and Nvidia in Intel fabs. Intel Products will probably sprint away from Intel Foundry which is fine, none of those products are strategic. It's a bad, bad scenario but for $53 billion, with the numbers people throw around for AI, it's doable.

Not doing it soon will lead to Intel crashing in about 12 months. I think their debt will be junk soon. So they're starting to fail as of August 2025.
In war Intel will say F to AMD and Nvidia and will only sell Intel products though even that can't be produced if the war is for Taiwan cause the ODM/MotherBoard Partners are mostly TW as well.
As for Intel failing look at AMD they were in even worse situation even after such a strong AMD comeback they are still not able to take 50% Market Share from a weak and screwing around Intel.
 
Back
Top