Personally if I'm a company I says "yeah your discount is nice, but uh, you should be shipping us a brand new non-compromised CPU for free anyway". Especially if the "fix" for so many of it is to keep reducing the performance via things like disabling HT, I'd be strongly re-evaluating being a heavily Intel focused customer.
I think Ars had an article on some older exploit that seems like it would work on Optane (it basically exploits data stored on memory during a reboot, which the fix is to make sure its pushed to the storage and then flush the RAM/memory before reboot - which if the storage and memory are the same as Optane's big potential is, then flushing it might be more problematic).
TBH, the biggest market for this kind of cpu is cloud providers, and they already have major discount, and intel also sold it directly to them.
Intel also makes custom chips for them (I think that's actually one of the big motivators behind EMIB development). AMD can of course as well, but that will be something that takes a little while to develop. Intel was smart to start doing that (although I think they also kinda had no choice, as if they tried to stonewall it, the companies would have looked at any and all alternatives to get what they want which would have set Intel on a fast track to getting x86/64 on the road to its dominance ending).
ARM is almost non-existent at the moment and has other issues such as the switch from x86 to ARM which might frighten some companies so I'm not sure Intel has to offer anything here
OTOH, as we all know, Intel heavily subsidized companies that took Intel chips for tablets.
I was thinking about the historical RISC players in the server market (MIPS, POWER, SPARC, etc.). But after all they might have collapsed just because they were too expansive and didn't (couldn't) properly react to Intel prices, even official prices.
Yeah, considering Intel itself failed pretty spectacularly with Itanium, I'm not sure how much I'd contribute that to their shenanigans vs there being other reasons. Which Intel's subsidies probably did play some role, making things that were marginal to consider already.
Thats why serving a niche market or just a smaller segment is impossible. You are simply priced out. Legal or not. Its impossible to avoid. People is educated for many years and trained to use monopoly like behavior.
Thats why zen needs to be sold from laptops over desktop to servers. They cant just adress one market.
I'm not sure I'd call it a niche market though. Its the bread and butter for Intel (and was the key area that AMD made inroads back with Athlon 64). There's a reason why AMD was really targeting this market with their Zen design.
And that's also why the Zen design was so smart, it really was focused on servers, but also to be scaled through the range.
I don't think this necessarily plays out. Due to some technical difficulties, this generation of EYPC is having difficulties getting into the cloud space. It is still very much Intel there. Rome could be a very different story.
From my conversations at VMworld, this is more about the composed bundles vendors are selling. When customers are doing refreshes, they're not *just* buying a blade chassis, or a single server. They often have a cluster, or big application, and are buying servers and storage to fulfill that role. They're buying 16 nodes for HPE Composed Infrastructure (or Dell Composed Infrastructure). They're buying VxRail Hyperconverged racks.
For instance, it's no quiet secret that Cisco is eager to get more marketshare in the Hyperconverged space, and is aggressively discounting Hyperflex pods to the point where depending on the deal, they're pushing All-Flash for the price that you'd only get Hybrid from VMware vSAN or Nutanix.
Likewise AMD has been heavily pushing the 50% licensing savings angle on VMware and VDI infrastructure. Intel is not trying to counter the price of individual CPUs, but rather they're going "Ok, with your business discount, you'd be paying $10,000 more in vSphere licensing with our solution, what if we just absorb that cost."
I'm definitely not a fan of this, but it's a very common practice in the industry.
I'd guess a lot of that is that many of the major cloud players get custom chips from Intel, and just generally that they're going to be a little hesitant on a new platform. I think we'll see that start to shift a fair amount as AMD is very willing to accommodate custom work, and the general situation (all the Intel security issues, Intel's pricing and production issues, AMD having a very competitive product).
Absolutely there's more to this than just the chips, which is another big factor in trying to get in. AMD I'm sure knew this, and knew they weren't going to make huge inroads just off the bat, but I think that's why they skipped the iterative change for EPYC and moved to it being their lead platform.