News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products

[511]

Then explain why they are investing a paltry 5B into Intel instead of 100B+. Riddle me that.

100B is likely to invite Monopoly scrutiny

 

[adroc_thurston]

100B is likely to invite Monopoly scrutiny

you just don't need 100B gigadollars to hijack a roadmap.

 

[Joe NYC]

You are one of the cool kids on the block that can NVLink with nVidia chip. What if nVidia says that it due to whatever it's NVLink or nothing? Then Intel is the only x86 vendor that can pair with nVidia. It basically cuts AMD from the market. Seems improbable but who knows. Unless AMD makes similar kind of deal. But then what will happen to its own GPU efforts.

The closest server CPU on Intel roadmap is Coral Rapids expected in 2028 / 2029. A lot of the things in the data center market place (CPU and GPU) will be decided before then.

AMD may be so far ahead of Intel by then in CPU, and sufficiently established in GPU that this will make no difference.

 

[adroc_thurston]

The closest server CPU on Intel roadmap is Coral Rapids expected in 2028 / 2029. A lot of the things in the data center market place (CPU and GPU) will be decided before then.

this can be a custom DMR too.

AMD may be so far ahead of Intel by then in CPU, and sufficiently established in GPU that this will make no difference.

yeah like you're not stopping them. Nothing ever could (or would).

 

[Josh128]

100B is likely to invite Monopoly scrutiny

Not under the current administration. I guarantee you Trump admin would have allowed a much larger investment than 5B if they were able to take credit for it. 100B would still be less than 50% of current valuation. Intel is an American icon, too big to fail, remember? An investment like that would virtually guarantee Intels competitiveness for the next 5 years and effectively remove all doubt that it could fail. That would be a 100% perceived win for Trump.

 

[adroc_thurston]

Not under the current administration. I guarantee you Trump admin would have allow a much larger investment than 5B if they were able to take credit for it. 100B would still be less than 50% of current valuation. Intel is an American icon, too big to fail, remember? An investment like that would virtually guarantee Intels competitiveness for the next 5 years and effectively remove all doubt that it could fail. That would be a 100% perceived win for Trump.

no one cares about the orange monkey, he'll croak to a heart attack soon enough anyway.
It's the product lines Intel and NV are collaborating on that are interesting.

 

[511]

you just don't need 100B gigadollars to hijack a roadmap.

You need a bunch of idiot to destroy a monopoly or hijack roadmap

 

[adroc_thurston]

You need a bunch of idiot to destroy a monopoly or hijack roadmap

It's honestly neither; just two companies helping each other in places where they are weak.
Now's the question: who's uke and who's seme here?

 

[511]

It's honestly neither; just two companies helping each other in places where they are weak.

I was talking about Intel destroying themselves

 

[adroc_thurston]

I was talking about Intel destroying themselves

Intel just couldn't lock in and IDC is washed.

 

[511]

Intel just couldn't lock in and IDC is washed.

what about the 10nm f*** up that is totally upto them and if they hadn't lost their lead their would be Foundry customer lining up

 

[adroc_thurston]

what about the 10nm f*** up that is totally upto them

Yeah that was stinky, but so was tying their IP and process roadmaps completely together.

 

[511]

Yeah that was stinky, but so was tying their IP and process roadmaps completely together.

TBH Everyone does this with TSMC if someday TSMC f*** up it will be a repeat. This cannot be avoided

 

[adroc_thurston]

TBH Everyone does this with TSMC if someday TSMC f*** up it will be a repeat. This cannot be avoided

oh noes.
People pick TSMC because they deliver the best nodes and they yield.
When they don't (hello N3b) everyone just kinda skips them.

 

[511]

oh noes.
People pick TSMC because they deliver the best nodes and they yield.
When they don't (hello N3b) everyone just kinda skips them.

well that doesn't change things if they did screwup everyone is screwed.

 

[adroc_thurston]

if they did screwup everyone is screwed.

Kind of? Only works under the assumption that SF can't perpetually yield.

 

[StinkyPinky]

I know a lot of people are seeing this as a bad thing and I'm wary, but there is also potential here. Intel need a shot of cash and they need Nvidia expertise. Nvidia is also rolling in money so I really hope they invest heavily into getting Intels fabs back to being level with TSMC at some point. It's not in our national interest (or Europe's) to rely so heavily on TSMC when they are literally right next to the PRC. China are rightly working hard to stop relying on foreign chip makers and we need to do the same.

 

[Josh128]

It's honestly neither; just two companies helping each other in places where they are weak.
Now's the question: who's uke and who's seme here?

Intel has nothing to lose from this, really. I wouldnt normally say they are the uke, but they are everyones uke at this time. They dont care.

 

[adroc_thurston]

I know a lot of people are seeing this as a bad thing and I'm wary, but there is also potential here. Intel need a shot of cash and they need Nvidia expertise. Nvidia is also rolling in money so I really hope they invest heavily into getting Intels fabs back to being level with TSMC at some point. It's not in our national interest (or Europe's) to rely so heavily on TSMC when they are literally right next to the PRC. China are rightly working hard to stop relying on foreign chip makers and we need to do the same.

well yeah the only real loser here are Tegra guys.

 

[gdansk]

Intel has nothing to lose from this, really. I wouldnt say they are the uke.

I think building a set of products around Nvidia GPUs that compete with it could cause considerable harm to Intel. If Nvidia prioritizes their own set of products, for example, it will be hard for Intel to compete when customers simply can't get their hands on the latest GPGPUs unless sold as part of an Nvidia rack.

It could also work fine. But risk remains. If the situation for Nvidia demands it why wouldn't they throw them under the bus.

 

[adroc_thurston]

it will be hard for Intel to compete when customers simply can't get their hands on the latest GPGPUs unless sold as part of an Nvidia rack.

wait what?
no. this isn't about that.

 

[gdansk]

wait what?
no. this isn't about that.

That's not the intent but plans can change based on execution and market situation.

 

[StinkyPinky]

I fully expect Nvidia to start investing heavily into Intel's labs and start making some of their products there in the 2030s. This is a win for Nvidia, they need to be diverse in their manufacturing range and again I think this is a win for the US and its allies. I personally think this is a win for Intel also, they need this shot in the arm. There are geopolitical concerns at hand also, it's not a coincidence that the Nvidia ceo Huang was in London yesterday with Trump. This is good news imo.

 

[Joe NYC]

this can be a custom DMR too.

Could be. But if DMR is H2 2026 +/- 6 months means H1 2027, and it is probably already taped out, you can add 1 year to that and it will be H1 2028

yeah like you're not stopping them. Nothing ever could (or would).

What are the odds that Intel comes back from being behind in CPU design and from being behind in process technology?

At least some client CPUs use TSMC, but server is only using Intel process technology.

 

[adroc_thurston]

and it is probably already taped out, you can add 1 year to that and it will be H1 2028

eh good enough.
MI500 comp here we gooo.

Intel comes back from being behind in CPU design

low.

from being behind in process technology

high.