Well no, the memory price increase is specifically here to price gouge AI datacenters. If you’re running a non-AI datacenter, and if you’re targeting consumers, like with remote desktop gaming, you have AI-like costs (literally, since servers GPUs are gaming GPUs), but without the enterprise customers. Oh and you also need the best bandwidth possible, because a few added MS might make the “remote” use unbearable, which is an added cost datacenters don’t have to deal with on the same terms.
Well no, the memory price increase is specifically here to price gouge AI datacenters. If you’re running a non-AI datacenter, and if you’re targeting consumers, like with remote desktop gaming, you have AI-like costs (literally, since servers GPUs are gaming GPUs), but without the enterprise customers. Oh and you also need the best bandwidth possible, because a few added MS might make the “remote” use unbearable, which is an added cost datacenters don’t have to deal with on the same terms.
That is prety irrelevant since consumers aren’t the customers for AI datacenter hardware, rich megacorporations with unlimited VC capital are.