If accurate, that would seem an encouraging sign of Apple’s long-term commitments to AI and Siri.
and historically Apple hates Nvidia
Interesting that with all the new M3 Ultra CPUs Apple is not doing Apple Silicon on their AI servers.
My explanations, Apple is in a hurry for a production-ready AI stack (I understand Nvidia has the lead on software) and/or they can’t build enough Apple Silicon processors for their consumers and their services at the same time.
I thought Apple’s Private Cloud Compute is being created so its M powered servers can handle user requests that cannot be fulfilled on-device?
It will be interesting to learn what Apple has planned for these Nvidia servers. The 2 hour Nvidia Keynote has already been viewed 37 million times in the past week.
Yes, I had that impression too. Perhaps this oomph is not for inference but for training custom models with proprietary capabilities. Who knows.
That 1 billion is pocket money and is not a serious investment in AI compared to other top players. Microsoft and Meta are both investing 60 - 80 billion this year.
As a boomer, let me say, I never expected to hear anyone say that about $1B.
It looks like these servers are being built by Dell, Asus, Supermicro, and . . .?
And, according to Asus’s Co-CEO it could take 6 months or more before customers should expect to start receiving them. So, IMO, they must be needed for some future project that has nothing to do with “AI Siri”. At least not for some years.
Nvidia’s GB300 will aid ‘very healthy’ growth in AI server business, says ASUS Co-CEO
I’m sure this is just to augment capabilities in the short term.