Apple Caught Off Guard by AI Demand for Mac Mini and Mac Studio
- Reference: 0185329704
- News link: https://hardware.slashdot.org/story/26/08/31/2133202/apple-caught-off-guard-by-ai-demand-for-mac-mini-and-mac-studio
- Source link:
> Apple normally releases new Mac models in the autumn, closer to October or November, making this week's announcement unusually early, falling just before the anticipated arrival of new iPhone models. The Information says that the AI-driven boom in Mac Studio and Mac mini sales is behind the early launch. Apple noticeably promoted the ability to link multiple Mac Studios together into a single, more capable system for running large frontier AI models, a feature aimed at business and developer customers rather than everyday consumers.
>
> Apple highlighted the Mac mini and Mac Studio's shift toward business buyers in June, with a "Business at the Park" event involving executives from major companies Ford, Disney, and Anthropic. The Mac mini was said to be the "darling" of the event. Even so, enterprise's rush toward powerful desktop Macs more broadly took Apple by surprise.
>
> The company reportedly did not possess an engineering team dedicated to business customers or staff focused on developer relations, and lacked an enterprise AI strategy. Businesses that approached Apple asking to buy access to the company's Private Cloud Compute infrastructure were reportedly turned down. Apple is instead leaning on partners such as WebAI and Mount Thor, which provide AI tools and execution environments built on Apple hardware.
[1] https://www.theinformation.com/articles/apple-stumbled-ai-hardware-success-mac
[2] https://hardware.slashdot.org/story/26/08/25/1553245/apple-announces-new-mac-mini-with-m6-and-m5-pro-chips
[3] https://www.macrumors.com/2026/08/30/apple-unexpected-mac-mini-and-studio-demand/
come on ... (Score:2)
... on the product page apple explicitly advertised how you can run large LLM on that hardware. They knew who's having the demand.
Of course (Score:2)
It's enterprises that don't actually understand what they are doing and trying to poorly shoehorn AI into their already shitty products.
We used to mock this (Score:3)
Integrated video cards had generally quite poor performance, but in the world of AI being able to share large amount of system RAM with the GPU make the MacMini a bargain for running your own AI workloads. For the cost of an RTX5090 you can get a MacMini with 64GB of shared RAM.
Re: (Score:2)
They don't offer their 128 GB versions anymore do they? For the good models, 128 GB is kind of the sweet spot. 256 would be ideal. Back when they offered 128 GB models, they were $10k.
Currently sort of looking at a computer with the AMD Rizen AI Mac 395 with 128 GB of RAM. Price is slowly climbing. When I last checked a few months ago it was a solid $3k. $3k pays for a fair amount of Claude Pro which is all I use right now. I would love to play with local models though.
I've been watching the youtube c
Who's money is being spent here? (Score:3)
Because OpenAI isn't even close to profitable, is Microsoft still giving them money, is Microsoft paying Apple to run unprofitable AI workloads now?
Its about running locally, not a data center (Score:5, Informative)
> Because OpenAI isn't even close to profitable, is Microsoft still giving them money, is Microsoft paying Apple to run unprofitable AI workloads now?
You misunderstand. While Apple does offer to run AI workloads in their data center rather than someone else's data center, suggesting greater privacy. The real offering in question is the ability to run AI workloads locally, on device, to not have to send potentially private information to a data center at all. Think of accountants and law firms that do not want data to leave their internal network. Software developers too. And various other's concerned about privacy.
A half truth is a lie, Sir (Score:1)
Even if you are running AI locally, if you are connected to the internet your information is up for grabs for AI training. It is well known that scrapers go behind [1]paywalls [imperva.com], [2]firewalls [substack.com] and general [3]hacking [bbc.com] of apps to get new data for training. Even if they are [4]blocked by IP [medium.com].
It is important to remember that when an AI agent hacks into a resource, it is reflecting the owners and their behavior. This should be about as surprising as learning that the son of a drug dealer is also a drug dealer.
I'm sure that thes
[1] https://www.imperva.com/blog/the-new-york-times-vs-openai-a-turning-point-for-web-scraping/
[2] https://mckayjohns.substack.com/p/web-scraping-hacks-that-changed-my
[3] https://www.bbc.com/news/articles/cp30989ee1wo
[4] https://medium.com/@Mustafa77/how-to-bypass-cloudflare-for-web-scarping-6e04839114d6
Re: (Score:3)
If the only thing you know about AI is what you read on a couple of headlines, then you may think just that. In the meantime there are truly countless local models including open source ones that have nothing at all to do with internet and simply being connected to the internet doesn't magically give up your data for training to anyone.
You seem to be confusing running a model from a major AI firm vs actually running one of the countless models yourself. AI and its use predates OpenAI, Anthropic and that ent
Re: (Score:2)
That's definitely hysteria.
Yes, if it's out there on the internet, AI can scrape it and it can be used for training. And there's only so much you can do about it.
That has nothing to do with being able to run AI on your own network, on your own system. You can run AI locally on an airgapped PC. Your data is as safe as the physical integrity of your PC or network.
Re: (Score:2)
Both the agent and the LLM model are running locally. There is nowhere to send the data for harvesting or training. That's the beauty of the local models. That's also the downside. They are frozen in time, their knowledge set at the time of original training. You have to download new versions every so often to get the latest updates, latest training.
Re: (Score:1)
Companies that want to self-host LLMs. If you're a heavy LLM user, a fleet of these running one of the Chinese open weight models is cheaper in the long run than paying OpenAI or Anthropic for tokens.
Re: Who's money is being spent here? (Score:2)
That's deluded bullshit. China produces items at all quality levels.
Re: (Score:2)
I have yet to see, or experience, anything from China that either: does not perform like shit or is not built like shit (or both). Now, to be fair, this is also true of a significant number of companies regardless of where they manufacture or are headquartered.
As someone who tests and uses a lot of AI models, the Chinese models do not impress me even a little.
Re: Who's money is being spent here? (Score:4, Insightful)
The rest of the world is driving Chinese EVs that are cheaper and rival the quality of all the legacy automakers. And before you open your mouth about safety, yes they meet the EU safety standards which are more stringent than the USA. China literally cranks out 10x the amount of engineers as this country and you think they are incapable of quality?
Re: (Score:2)
> The rest of the world is driving Chinese EVs that are cheaper and rival the quality of all the legacy automakers. And before you open your mouth about safety, yes they meet the EU safety standards which are more stringent than the USA.
Great, I hope to be able to experience at least one at some point to judge them specifically. I'm all for getting these shitty oil burners off the road. However, the reason they are cheap is because the Chinese government is subsidizing the manufacturers, so I don't expect that to always be the case. The range of the EVs needs to be tested locally in realistic local road conditions so that the range is not just a theoretical max that is never reached in daily real-world scenarios. Which means the ranges of
Re: (Score:2)
The Chinese goverment doesn't subsidize them that much.
Re: (Score:2)
Haha! You have been sleeping under a rock for the last 10 years. Phone and laptop chargers, cables, shoes, woodworking tools, clothes. All that in almost daily use and not thrown out, not lost colour or parts for 5 years. Try that for locally bought similar but twice as expensive products. 2-3 years and unusable.
Re: Who's money is being spent here? (Score:1)
> I have yet to see, or experience, anything from China that either: does not perform like shit or is not built like shit (or both).
They have a pretty successful [1]space program. [cnsa.gov.cn]
[1] https://www.cnsa.gov.cn/english/
Re: Who's money is being spent here? (Score:1)
lol what? There is a microsoft ghost behind every closed door on in your diseased mind
Re: Who's money is being spent here? (Score:1)
If Microsoft cuts a check to OpenAI or whoever, the money stops being Microsoft as soon as OpenAI cashes the check.
When grandma sends you a $5 birthday check, it isn't grandma's money your spending at the comic book store, it's yours.