AMD Instinct MI300X Released at Opportune Moment. NVIDIA AI GPUs in Short Supply

[T0@st]

LaminiAI appeared to be one of the first customers to receive an initial shipment of AMD's Instinct MI300X accelerators, as disclosed by their CEO posting about functioning hardware on social media late last week. A recent Taiwan Economic Daily article states that the "MI300X is rumored to have begun supply"—we are not sure about why they have adopted a semi-secretive tone in their news piece, but a couple of anonymous sources are cited. A person familiar with supply chains in Taiwan divulged that: "(they have) been receiving AMD MI300X chips one after another...due to the huge shortage of NVIDIA AI chips, the arrival of new AMD products is really a timely rainfall." Favorable industry analysis (from earlier this month) has placed Team Red in a position of strength, due to growing interest in their very performant flagship AI accelerator.

The secrecy seems to lie in Team Red's negotiation strategies in Taiwan—the news piece alleges that big manufacturers in the region have been courted. AMD has been aggressive in a push to: "cooperate and seize AI business opportunities, with GIGABYTE taking the lead and attracting the most attention. Not only was GIGABYTE the first to obtain a partnership with AMD's MI300A chip, which had previously been mass-produced, but GIGABYTE was also one of the few Taiwanese manufacturers included in AMD's first batch of MI300X partners." GIGABYTE is expected to release two new "G593" product lines of server hardware later this year, based on combinations of AMD's Instinct MI300X accelerator and EPYC 9004 series processors.



View at TechPowerUp Main Site | Source

 

[Pumper]

It's not really a great look when your product sells only because the competition can't meed demand.

 

[remekra]

Yes but does it have DLSS and Ray Tracing?

 

[Vya Domus]

It's not really a great look when your product sells only because the competition can't meed demand.

These things have probably been ordered a long time ago and I bet the main reason companies choose AMD simply has to do with the fact that their products are probably several times more cost effective.

 

[AnotherReader]

It's not really a great look when your product sells only because the competition can't meed demand.

When you're seeking a foothold in a market dominated by one player, you'll take anything. It doesn't matter how they get it, but AMD needs their product in the hands of customers. The hardware is pretty good, but they need the software ecosystem and that won't happen without customer adoption. Nvidia's biggest advantage isn't hardware, but the CUDA ecosystem.

 

[Tomorrow]

Yes but does it have DLSS and Ray Tracing?

These are server products we are talking about here. These things dont even display outputs. Much less upscaling.
Tho i suppose there is some use for RT and Tensor cores...

 

[thesmokingman]

It's not really a great look when your product sells only because the competition can't meed demand.

Don't know much about these do ya? There are dozens of AI chips out there, the problem is none other than Nvidia and AMD can produce in volume. AMD's gpus have a 60% uplift and AT HALF THE COST, that too must not be a great look huh?

 

[Dave65]

Don't know much about these do ya? There are dozens of AI chips out there, the problem is none other than Nvidia and AMD can produce in volume. AMD's gpus have a 60% uplift and AT HALF THE COST, that too must not be a great look huh?

Obvious he is a troll.

 

[JAB Creations]

It's not really a great look when your product sells only because the competition can't meed demand.

How long have you been working for Nvidia?

 

[Wirko]

When you're seeking a foothold in a market dominated by one player, you'll take anything. It doesn't matter how they get it, but AMD needs their product in the hands of customers. The hardware is pretty good, but they need the software ecosystem and that won't happen without customer adoption. Nvidia's biggest advantage isn't hardware, but the CUDA ecosystem.

AMD also needs precious wafers out of hands of Nvidia. It seems they are having enough success doing that. I have no idea how bidding for manufacturing capacity looks like, but you probably get more if you pay farther in advance.

 

[LabRat 891]

I recall researching the MI300(X) when I was buying up MI25 Vega10 cards.
Extraordinary things, the MI300s are.

-Massive MCM arrangement with an insane amount of CDNA cores, and HBM stacks.

 

[Minus Infinity]

When you're seeking a foothold in a market dominated by one player, you'll take anything. It doesn't matter how they get it, but AMD needs their product in the hands of customers. The hardware is pretty good, but they need the software ecosystem and that won't happen without customer adoption. Nvidia's biggest advantage isn't hardware, but the CUDA ecosystem.

In AI CUDA's role is waning. People are eager for hardware agnostic software. ROCm is doing well and we have OpenAI, Pytorch etc that can run on any hardware and are making big inroads. CUDA might be king in other areas for a while but not in AI for too long.

 

[JohH]

It's not really a great look when your product sells only because the competition can't meed demand.

I don't think they care about how it looks. Selling parts = good. Having an install base which will lead to further contributions in open source projects = good.
It's also basically the origin of AMD's x86 processors. They were merely a second source for Intel.

 

[LabRat 891]

It's also basically the origin of AMD's x86 processors. They were merely a second source for Intel.

I was about to say...
It doesn't matter "how it looks"
Many companies around today started out (or made in-roads into their market) in that manner.