• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

BloombergIntel in talks with AMD to license GPU patents

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
This deal would effectively kill off AMD's APU market which is the only reason they still have a CPU division. Unless this is part of the graphics division's plan to secede from AMD without AMD's knowledge, this doesn't really make any sense for them. With NVidia not having any x86 CPU's and Intel not selling dGPU's or anything above the really low end, the deal makes more sense for Nvidia since there are few areas where the 2 companies are direct competitors. If the rumor is actually true, the most likely scenario is that AMD and Intel (independently and for their own reasons) are just throwing up a smokescreen and trying to decrease the money that Nvidia gets from an agreement renewal.
 
They are talking "IP" in the same sense that the NVidia deal covered; patent cover to stop your ass getting sued. They didn't just shove NVidia GPUs into their chips, and they're not going to shove AMD GPUs into their chips. This is about patents, not specific designs.

Seems like a sensible move for Intel; throw AMD a lifeline to keep anti-trust off your back, stop feeding money to NVidia (who seem like the more credible rival right now), and if AMD goes belly up you can just acquire the portfolio at a knock down price.
 
I would assume they are actually going to license IP from both NVidia and AMD to keep of restrictions. The amount Nvidia gets is peanuts with 22 million per month. AMD would get same or less.
 
ShintaiDK said:
I would assume they are actually going to license IP from both NVidia and AMD to keep of restrictions. The amount Nvidia gets is peanuts with 22 million per month. AMD would get same or less.
Click to expand...

Nvidia initially received more, with payments tapering off in the latter half of the license period. But even at the lower rate, an extra $50m per quarter would be equivalent to 5% of AMD's total revenue. Chump change to Intel, but AMD needs all the money that they can get their hands on right now.
 
NTMBK said:
Nvidia initially received more, with payments tapering off in the latter half of the license period. But even at the lower rate, an extra $50m per quarter would be equivalent to 5% of AMD's total revenue. Chump change to Intel, but AMD needs all the money that they can get their hands on right now.
Click to expand...

The deal was ~1.5B over ~5 years, or ~300M per year. Or ~25M per month that turned out to be ~22M$.

Its a fixed price all through the period.

Intel will pay NVIDIA $1.5 billion over the next 5 years. This obligation will be recognized as a liability totaling approximately $1.4 billion, on a discounted basis. Intel recognized an expense of $100 million in the fourth-quarter of 2010, classified as “marketing, general and administrative.” The remaining amount, approximately $1.3 billion, will be recognized as an intangible asset in the first quarter of 2011 and will be amortized into cost of sales over future periods
Click to expand...

A extra 50M$ per quarter for AMD would create a minor miracle on its own and possible put their finance in the black once a while instead of the red flow.
 
Last edited:
my feeling is this is a down the road play.

with google apple ms and every other heavyweight jumping into the AR market, at some point intel's IOT is going to need some serious graphic horsepower in order to process realtime overlays on realworld reference trackpoints sans pre-arranged IR or optical markers. all those soc that intel wants inside every consumer item will need some ability to reference spatial data to report to all the other soc.
 
gorobei said:
my feeling is this is a down the road play.

with google apple ms and every other heavyweight jumping into the AR market, at some point intel's IOT is going to need some serious graphic horsepower in order to process realtime overlays on realworld reference trackpoints sans pre-arranged IR or optical markers. all those soc that intel wants inside every consumer item will need some ability to reference spatial data to report to all the other soc.
Click to expand...

Good catch, I had not considered VR or AR at all.

AMD's tech is still the leading edge for these applications. If it's a potential big market, ofc Intel is going to want a piece of it.

And obviously, $66M per quarter is chump change for Intel, but it would be a good life-line for AMD's revenues.
 
90% sure they aren't licensing actual implementable technology. They just dont want the threat of lawsuit hanging above their product lines. I highly doubt Intel's GPU plans will change in any meaningful way as a result of this. The path is already plotted, this just makes sure that path is clear. Patents can be very broad, and the threat of a "reasonable royalty" on your entire product line makes the risk assessment extremely easy. It's a no brainer to license the patents first and largely mitigate the risk.
 
Headfoot said:
90% sure they aren't licensing actual implementable technology. They just dont want the threat of lawsuit hanging above their product lines
Click to expand...

/this but still doesn't make sense to me.

They don't care about AMD's tech, they just want to design their own gpu's without getting sued by AMD or Nvidia. However, I thought AMD already has a cross licensing agreement with Intel, due to them both developing x86 cpu's. Do they need another one for gpu's, if so why now as they have been developing gpu's for years?

They would still need to license nvidia's patents, AMD has a cross licensing agreement with nvidia so they can both develop gpu's but surely AMD can't extend that to license Intel to use Nvidia's gpu patents?
 
Dribble said:
/this but still doesn't make sense to me.

They don't care about AMD's tech, they just want to design their own gpu's without getting sued by AMD or Nvidia. However, I thought AMD already has a cross licensing agreement with Intel, due to them both developing x86 cpu's. Do they need another one for gpu's, if so why now as they have been developing gpu's for years?

They would still need to license nvidia's patents, AMD has a cross licensing agreement with nvidia so they can both develop gpu's but surely AMD can't extend that to license Intel to use Nvidia's gpu patents?
Click to expand...


Does Nvidia own AMD's IP or something? Why imply that? :sneaky:

Intel licensed AMD's AMD64 and AMD gets their coveted x86 rights which Nvidia at this point dying for.
 
AMD fanbois rejoice...

In reality, this is just another cost-cutting move by Intel. As previously pointed out, they'll likely see upwards of a 50% costs savings by licensing thru AMD. No, we won't see any significant Integrated Graphics changes from Intel. If you think otherwise, keep dreamin....
 
Erenhardt said:
Console $ was too little for nvidia to participate, you think a giant like intel would even considered such silly little project?
Click to expand...

Or so NVidia claimed. They were perfectly happy to accept console margins for two generations in a row, then when they didn't have something competitive to offer they claim they never wanted it in the first place.

Both companies wanted an integrated all-in-one chip, and they wanted a 64-bit CPU. NVidia weren't in a position to offer that so they were out.
 
Is this Intel providing life support to AMD so they can maintain a weak competitor and stay out of the FTC's sights? And/or with the dominance of ARM in devices, will anti-trust concerns diminish in the coming years and this is positioning by Intel to acquire AMD and consolidate the x86 market in the future?
 
thesmokingman said:
lol, just read Nvidia is releasing some of GW to open source. Probably out of the goodness of their heart right?
Click to expand...

They are feared to death about GPUOpen. That is why many things started to happen. Including "opening" CUDA. Especially when they saw that Boltzmann Initiative works, and HPC market is starting to turn over into AMD Direction.

End off-top.
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top