Ordered 9th, paid 24th, shipped 25th. Desktop did a trip from Taiwan to Paris, France, then Milano, Italy and back to Paris. Destination is Germany.
This is quite consistent with my usual Fedex experience. Them sending my packages on trips through Europe.
That’s hilarious. Mine did a couple of trips to Angles City in the Philippines, then Guangzhou before finally heading to Australia. I can only assume some gastronomical motive 
Yeah, mine was in Angeles City as well. Then China. I guess Fedex doesn’t go directly from Taiwan to China directly because of… reasons…
My guess is that there is some “smart” machine that puts it into a container (ULD) where there’s still space left. This container goes to City X. Gets unpacked there. Get’s directed to the proper destination the usual way (which sometimes is the same way back) and there it is hoping it gets unpacked and put in the correct ULD.
There’s probably no human contact anywhere.
It’s probably cheaper/faster to have those wrong flights than checking the proper destination of everything in a container.
I guess I should update this- I got my Framework Desktop last Tuesday. I have been mostly happy with it, though not without some major troubleshooting- none of it is Framework’s fault.
First, I wanted to use Fedora KDE, but encountered lots of problems with compatibility with my software (no Linux option, so I spent a lot of time trying to get alternative solutions, which failed) and some projects I’m working on. Eventually, I just decided it wasn’t worth the trouble and saw a Black Friday deal to buy a copy of Windows 11 Pro for $10 so I jumped on it. I wanted to go with Linux because of the extra few VRAM it could allow, but I didn’t really see that difference anyway. I could only set the VRAM allocation to 96 gb, which is the same as on Windows. Plus, with the headaches with software and setup, I bailed. I might try again later on, but I have a lot of work to do with starting a company, and I don’t need the added complication of figuring out every step.
The other challenges were hardware compatibility-related- I wanted to run a cluster with my old pc, but decided to turn my Radeon RX7900 XTX into an eGPU for 128+24 GB VRAM with some cheap hardware and struggled to get it to play nice with the new system. I finally did, but it took some time. I’m still finetuning this.
Lastly, something I am still having problems with is that I wanted to use my PCIe slot for an additional NVMe SSD, but have not found a compatible adapter. I bought two different ones from Amazon, but neither has worked. I’d love some suggestions from the group of one that works with this system, as this is something I still want to do.
I currently have this one in my framework desktop: Amazon.com: GLOTRENDS PA09-HS M.2 NVMe to PCIe 4.0 X4 Adapter with M.2 Heatsink for M.2 NVMe SSD : Electronics
I use it it with a Samsung 990 pro 1tb running Fedora 43 workstation.
Which tech did you use to have egpu? Oculink?
And how does it run?
I’m running a Radeon RX 7900 xtx via USB4. So far it seems pretty good. No major issues, other than I had to abandon LM Studio and switch to Ollama so it would recognize both GPUs when loading models. USB4 isn’t quite as fast as Oculink, but that speed of transmission isn’t really important when running LLMs. I’m still dialing in how to optimize it, as I have been preoccupied with some other things. I’ll check back in a few weeks after I’ve had a chance to give it a good evaluation, but so far, so good.
Thanks for the feedback, good luck with your concern 
There’s a long thread on the topic of dGPU’s here @Ernie_Hobbs It might be helpful to keep these topics together as the Batch threads don’t really help others, who might search on the topic, find useful information like yours.
Cheers!
Order date Nov 04 2025
Delivery Batch 15 of Framework Desktop Edition (AMD Ryzen™ AI Max 300 Series) - 128GB
Delivered · Delivery date: Dec 5, 2025 (Fri)
