Many products are partially assembled elsewhere and final assembly in America. I’m good with what Apple is claiming.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple Starts Shipping Made-in-America AI Servers Early
- Thread starter MacRumors
- Start date
- Sort by reaction score
Apple has yet to deliver the Apple Intelligence features it previewed at WWDC 2024, that were originally to be released with iOS 18. No real improvements with Siri, in fact Siri has become worse in some basic features. I personally don’t find Siri useful at all, and it absolutely makes no sense to use Siri as a middleman to access ChatGPT. Apple is definitely behind the game in basic AI use with a personal assistant such as Siri in comparison to Google and others.
This. Apple currently has a decent moat around its ecosystem. They basically have about two years to figure out how to add enough AI to its ecosystem to prevent someone like OpenAI from developing a competing OS/ecosystem/hardware portfolio that will meaningfully pull market share away.
Apple would be such a better company if Steve Jobs was still living, or if they continued to follow his vision and principles. The core of the Apple is slowly rotting ….
sure, they're behind on it, but it doesn't answer the question:
how is this a mortal wound that Apple don't have time to fix?
What is this precipice they are so in danger of falling off of?
So much electricity and water wasted on AI slops!
Yeah, and then when Apple offers a free PRIVATE LLM (emphasis on Large) via the cloud, via your iCloud (family or otherwise) Subscription. Where all your queries, caching, and history, are 100% private, people will go whoa, hmm this is something.
Because right now the closest thing to trying to be private is Ollama.app which is hampered, limited, still costs $$$, and still not 100% private. i.e. They are working on the future of AI where privacy first is paramount. Companies will pay to have this, and not set up their own data centers.
Because right now the closest thing to trying to be private is Ollama.app which is hampered, limited, still costs $$$, and still not 100% private. i.e. They are working on the future of AI where privacy first is paramount. Companies will pay to have this, and not set up their own data centers.
Please do not conflate us (Europe) with whatever the US has become.
Whatever the western values were, don't exist anymore in the US of Trump.
Ollama is 100% private, running in your hardware, while you then went on to describe handing everything off to Apple’s servers… as if that was an improvement??..
Does it matter? It still earns them brownie points. Let's tale this one step at a time.
I think it’s safe to assume the silicon is M2-Ultra based and comes from Fab 21 (Phase 1, 5nm/4nm) in Arizona. The chassis is probably made in Houston (or Austin), but most other components are imported.
So these are replacement/expansion servers for existing PCC data centers.
We can speculate about what the next-generation PCC servers might be, maybe a specialized M5 variant that could be produced at Fab 21 (Phase 2, 3nm) when it begins volume production in 2028. So it would follow the same pattern, with the PCC data centers that use the next-generation servers already in existence by then.
Last edited:
The only thing that matters to me is how energy efficient Apple can make its AI servers and how that translates into performance per watt. At some point AI is going to hit a tipping point from a power consumption cost and politicians are going to get involved and/or the general public is going to start protesting this en masse. The current power cost/benefit is ridiculously bad at this point. I need solid, resilient infrastructure for my power, not puppies and cats flying jets videos.
Apple is going to defer that to the EU at this point. The EU seems to be in control of its own technological progress whether you think tech companies are being petty and spiteful is entirely based on your personal point of view.
And even in the case of OpenAI making hardware…well…
OpenAI and Jony Ive Reportedly Struggling to Define Their AI Device Vision - AppleMagazine
OpenAI’s highly anticipated collaboration with legendary designer Jony Ive appears to be running into turbulence. Sources close to the project say the pair are still struggling to define what the...
applemagazine.com
The casing on these units is a simply lovely shade of Naranja Amarga. The color represents the beauty and power of the richly vibrant Apple Intelligence ecosystem.
That’s would cost a lot more than $600B, and would take at least a decade or more