None of that will do anything. If the massive TCO and perf/$ advantages AMD offer are not enough to sway some companies from Intel nothing will. It is not a fight AMD can win on pure economics so why even try?
Also if businesses want to saddle themselves with a higher TCO and lower performance part they will get out competed by those that go AMD, it just takes a while for those kinds of decisions to bite you.
Then there is timing of payments Vs timing of build for the super computers. Pretty sure AMD are getting the payment next quarter but they have been building the parts for a while which eats wafer supply possibly making it pretty hard to offer better prices without inducing a shortage.
So while in theory AMD could dump prices if they don't have the wafer supply to have sufficient product at that price point then they won't exactly sell many more products but will make a good chunk less in the process.
Also as down as Intel are they would win a price war at a cost and NV would dominate one. It is not something AMD can win in the long term because they still do not command enough of the market to make a volume play.
When AMD lost 66% of sales, AMD had the opposite problem. Chips sitting in inventory. Not a problem delivering volume of chips.
And, before the bottom fell from below the market (and Intel aggravated it by anti-competitive shenanigans), AMD had a plan to grow the volume of wafers from TSMC.
So when AMD client sales were 2.1 billion, it was positioning itself to grow to 2.5 billion. And instead of selling 2.5 billion, AMD sold ~700 million.
If anything, AMD had to pay penalties, or negotiate a deferral, order cancellations with TSMC to go from planned 2.5 billion down to actual ~700 million.
BTW, if AMD did not have super profitable Xilinx and X-Box divisions, that anticompetitive blitzkrieg Intel executed on AMD would have left the old AMD without Xilinx on the verge of bankruptcy, laying off staff, closing offices.
This is why I have consistently said AMD need 50% market share in most of the sectors to really be secure because if AMD were to miss and Intel were to hit one out of the park the market share AMD have gained over the last 5 or so years will disappear overnight. A surefire way to have a miss is to enter into a price war and then not have sufficient money to support the R&D. Pretty much the only sector AMD are secure in is console APUs. Every other one will get eaten up if Intel have anorher Conroe moment.
Yup. And the paradoxical thing is that AMD let Intel get away with it when AMD is in far better financial position and Intel was in its weakest position. Just threatening to match Intel prices and tactics would send shivers down Pat's spine and he would back off.
It's like AMD is playing checkers, Pat playing chess.
Or Pat being dealt a weak hand in card game and winning decisively.
This industry seems almost hopeless if Intel is getting away with this, if AMD is losing market share while having much stronger products. And lower costs.
The underutilized Intel Fabs are an albatross weighing on Intel. And instead of emptying them further, AMD filled them up, lowering Inte's underutilization charges.
It's almost like a Chips Act by AMD, supporting Intel fabs...
So no, your strategy would work to bolster revenue in the short term but divert too many resources from R&D to operations along with the massive headcount increase that would require in a short period which will mean hiring lower quality candidates and then it all spirals down. NV did not get into their position by price warring, they got there by steady consistent execution over decades. The last fail hardware they released was probably the GTX 480 and even that fail was still the fastest, it just came with a power cost and had the George foreman grill memes. Since then every arch has been a hit, NV are a reliable execution machine and that is what AMD need to do and really have done on the CPU side since Zen. Not so much on the GPU side but even then they are executing well on the HPC side it is just consumer is a bit more up and down.
I don't think price war is the only weapon. There are other strategies. Hire an army of sales people. Engage in marketing / advertising campaigns. Support the people who want to buy and use your products. Big and small. Small could become big - as could happen with AM5 micro servers, for example, if AMD gave it some support.
To get a picture of AMD's lack of effort, AMD's own Zen 4 drivers have no support for Windows Server. I fiddled with it and got the drivers to load. So it is bare minimum that AMD is not doing to try to sell good chips, while Intel is doing everything to sell mediocre chips.
I am not complaining about AMD market share in GPU, because it is what it is as a result of mixed level of execution.
CPU is a different story. AMD had consistent execution, has been 50:50 plus / minus vs. Intel in desirability and performance of the products. Even more consistent than Intel.
And all that brought AMD is 1:3 disadvantage in unit sold, and from that pathetic state, a further 66% drop in revenue.
Intel is seen as essential, and AMD allows itself to be seen as non-essential vendor in client PC hardware. This gives Intel power over OEMs, to strong arm them into compliance. This is something AMD needs to break to be viable in client PCs.
And it is one of the few things
@adroc_thurston is completely wrong about.
github.com
www.phoronix.com
forums.anandtech.com
Facebook
Twitter