Thursday, November 28th 2024
Intel's $7.86 Billion CHIPS Act Grant Forbids Selling Its Foundry Business
When Intel announced the completion of its $7.86 billion CHIPS Act grant from the Biden-Harris administration on Tuesday, we assumed some special terms were tied to the grant. Intel is essentially making a law-abiding promise to the US government that it will not sell its stake in the Intel Foundry unit under any circumstances, even if it manages to become an independent entity. This ensures that Intel is the major voting party in any event. Intel disclosed in a regulatory document that if Intel Foundry becomes its own private entity, Intel must maintain majority control with at least 50.1% ownership to keep its subsidy agreements. Additionally, if Intel Foundry goes public in the future, no single investor would be allowed to acquire more than 35% of shares unless Intel remains the largest shareholder, as this would trigger control-change clauses.
This essentially positions Intel Foundry as too big and too important of a unit to fail, both for Intel and the US government. Given Intel's ties with the US Department of Defense, with up to $3 billion in direct funding under the CHIPS and Science Act for the Secure Enclave program, Intel is vital for providing the US government with advanced semiconductor manufacturing. Strategically, Intel Foundry is the sole US-based company that competes with advanced manufacturing companies such as TSMC and Samsung. Even with TSMC and Samsung driving investments on US soil with advanced fabs, Intel's work with the government requires additional safety and secrecy clearances that only a US firm could provide. In the latest Q3 2024 financial results, Intel Foundry recorded a revenue of $4.4 billion with $5.8 billion in losses. While the operating marking of negative 134.3% seems like a disaster, upcoming quarters will bring it to a positive with more customers and using already developed nodes like 18A.
Source:
Reuters
This essentially positions Intel Foundry as too big and too important of a unit to fail, both for Intel and the US government. Given Intel's ties with the US Department of Defense, with up to $3 billion in direct funding under the CHIPS and Science Act for the Secure Enclave program, Intel is vital for providing the US government with advanced semiconductor manufacturing. Strategically, Intel Foundry is the sole US-based company that competes with advanced manufacturing companies such as TSMC and Samsung. Even with TSMC and Samsung driving investments on US soil with advanced fabs, Intel's work with the government requires additional safety and secrecy clearances that only a US firm could provide. In the latest Q3 2024 financial results, Intel Foundry recorded a revenue of $4.4 billion with $5.8 billion in losses. While the operating marking of negative 134.3% seems like a disaster, upcoming quarters will bring it to a positive with more customers and using already developed nodes like 18A.
18 Comments on Intel's $7.86 Billion CHIPS Act Grant Forbids Selling Its Foundry Business
We only keep getting bold statements by Intel's staff and "We shall prevail in 2025, trust me" by Patty Promise-singer aka Pat Gelsinger.
Take this two above quoted senteces not with a bit of salt, but more like with tons of it. Submerge yourself in the salt already!
The only way to work with the requirement is for Intel to stop making it’s own chips. The incompetence of both Intel and the US government means these two are a match made in heaven.
www.techpowerup.com/329264/qualcomm-abandons-intel-takeover-focuses-on-division-acquisitions
Given your irritation towards an Intel acquisition, which camp are you in: people who love Intel so much and can’t stomach the fact that Intel has fallen so badly that they could be bought or people who realize Intel is doing so badly that no one in their right mind would buy them?
The biggest issue right now is Intel's (x86) fixed at the hips with IFS; but that needs to change completely for them to recover!
AMD isn't selling more chips because they bake them on a better node. They sell more chips because they develop better chips. And they sell more chips because its much easier and cheaper to make them, and thus price them competitively.
Nothing here has to do with node advantage or disadvantage. We've seen Nvidia beat AMD GPUs on bigger nodes too. And not by a small margin. All I see here is what was clear from the moment AMD moved to chiplet and got functional CPUs out of it: they will get better yields because they can keep making smaller chips, and they have a scalable architecture. Intel does not, keeps pushing the monolithic button, and even now they haven't moved to chiplets proper, its a wild glue fest of separate chips from different nodes stuck to a substrate. Its not properly scalable without redesigning the floor plan entirely, or making some chips bigger over time, which is exactly not what AMD is still advancing on. Intel is on an architectural dead end yet they keep adding new miles to the road, hoping they might somehow find an end to the dead end, or something. Big little is of the same nature; it exists only because their big core is too hungry and won't scale without destroying yields.
What's killing Intel right now is that they have a competitor that has already adjusted to the new paradigm in chip production: smaller is better, because big is expensive and risky, which cascades into having trouble securing foundry capacity, or building towards new nodes, because how are you going to get good yields with these massive chips? That is why we saw Ultra move towards a wild mosaic of different sized chips as it has right now, sourced from all over the place to arrive at a somewhat functional product.IF you are of the opinion that the market is the most important thing in the world.
But clearly, its not. You need the market to drive economy... but when people start shooting, we've seen how quickly the market evaporates entirely.
I think the US just secured its chip production going forward, and IFS... nothing of value was lost IMHO. That shit was never gonna work.