Apple’s new Mac Mini and Mac Studio Are All About Local AI - DTNS 5339 artwork

Apple’s new Mac Mini and Mac Studio Are All About Local AI - DTNS 5339

Daily Tech News Show

August 25, 2026

Uber is giving parents of teens the ability to livestream video of the child's ride in real time, and Android 17 includes a new Motion Assist feature to help reduce motion sickness for riders in moving vehicles. Starring Jason Howell, Tom Merritt and Dr. Niki.
Speakers: Jason Howell, Tom Merritt, Dr. Niki

Topics: Technology, News

**Jason Howell** (0:04)
This is the Daily Tech News for Tuesday, August 25th, 2026 We tell you what you need to know, give you the important context on all of it, and help each other understand.

**Tom Merritt** (0:13)
Whoa, today Dr. Niki is going to explain why making female clones of male mice is important for you.

**Dr. Niki** (0:19)
Me?

**Tom Merritt** (0:20)
Don't say it's not.

**Jason Howell** (0:20)
Them.

**Tom Merritt** (0:22)
Okay. Yeah. Also, the big story is Apple has new Mac Minis and Mac Studios that are perfect if you want to run a local AI.

**Jason Howell** (0:30)
Oh, boy. The story of 2026 localized AI. I'm Jason Howell.

**Tom Merritt** (0:35)
I'm Tom Merritt.

**Jason Howell** (0:36)
Let's start with what you need to know with the big story.
Yes, Apple has announced new versions of the Mac Mini and Mac Studio and some new chips. We're going to talk about all of this. But really, it seems like the big story here is that what Apple is announcing is very much addressing developers and people who might want to run open weight AI models locally on their machine, right? The Mac Mini now runs on M6 or M5 Pro. Apple is claiming up to four times faster AI inference, twice the graphics performance from previous models as well. The M6 is, by the way, Apple's first two nanometer Mac chip. Has a 12 core CPU, 12 core GPU, up to 32 gigs of unified memory. 32, we'll talk about that. 160 gigs of memory bandwidth and starts at $899. Then you've got the M5 Pro model starting at $1699. Both of them, by the way, $100 more than the M4 model equivalents that came before them. Then you've got the Mac Studio, the beefier, bulkier one. It's been at the gym, let me tell you.
Gets M5 Max and M5 Ultra chips, 4.3 times faster inferencing. The M5 Max starts at $2499. The M5 Ultra starts at $5499, so $5500, both an increase of a couple of $100 from before. We've got the chips which we can talk about for sure, but why don't we just start with the machines themselves?
Maybe what this seems to be, which is Apple creating a hardware that could potentially probably will appeal to people who are looking at this moment in AI and realizing, hey, I don't want to be in the cloud for everything. I want to do this all locally. Here's some machines that are built kind of with that in mind.

**Tom Merritt** (2:33)
Yeah, you can daisy chain Macs to run inference on LLMs. And so this is saying, hey, wouldn't you like to have more daisy chain? More daisy in your chain, I guess? I don't know. We can talk a little more about why that is. When we talk about the chips. But a lot of it is made possible by Mac OS 26.2, which enabled some low latency communication between Thunderbolt 5 hosts. And so that continues here. And the increased chip power means that you can do more of that. It's also a great, the Mac Mini itself is also a great machine for people who are not at all interested in inferencing locally. For sure.
And I think a lot of people have been waiting for this.
The conventional wisdom is pretty much similar to every other release, which is, if you've got an M4 Mac Mini, you probably don't need to upgrade. You got an M2 or an M3 Mac Mini, now you might want to look into it, even though you are going to pay more. $100 more, maybe not as bad as a lot of people thought. Apple's been pretty good at holding the line on prices. So the fact that they had to raise it at all, I think is telling. But yeah, I think if you're on an M2 Mac Mini and it's starting to chug, this is a great option for you. If you're on an M4 and you don't need extra power for some kind of supercharged thing, then you probably don't need to upgrade to this. And the Mac Studio, I don't know, it's probably the same diagnosis I would give for that. But also the Mac Studio to me is for people who are silly.
I don't mean silly, like you're not a serious person, but yes, I'm not concerned about whether they should upgrade or not. I just want the maximum thing.

**Jason Howell** (4:19)
Whatever the biggest, most evious thing. Yeah, I mean, I'm running a Mac Studio right now to do this, so call me silly, Tom.

**Tom Merritt** (4:26)
You're a silly person. No, I just mean like who needs, I guess I shouldn't say you're silly. I should say you need silly power, right? To do the things that you're doing.

32 more minutes of transcript below

Thousands of transcripts fetched by people building searchable podcast archives

Feed this to your agent

Try it now — copy, paste, done:

curl -H "x-api-key: pt_demo" \
  https://spoken.md/transcripts/1000651996090

Works with Claude, ChatGPT, Cursor, and any agent that makes HTTP calls.

From $0.10 per transcript. No subscription. Credits never expire. Prices exclude VAT, added at checkout for EU customers. Not what you expected? Email us within 14 days with 20 or fewer credits used and we refund the pack in full.

Using your own key:

curl -H "x-api-key: YOUR_KEY" \
  https://spoken.md/transcripts/YOUR_EPISODE_ID