Five mainboards in eight months, refund offered and hardware returned, three weeks of silence

I have been going back and forth with support since January and I have run out of other things to try, so I am posting here.

The short version. I bought a Ryzen AI 300 mainboard upgrade in November 2025, around 1,770 euros for the mainboard, 2.8K display, top cover, battery, webcam, Wi-Fi card, hinges and speakers. The mainboard failed about eight weeks later. Since then I have been through five mainboards and one complete replacement laptop, and I have not had a working computer since March. In July support offered me a refund, I accepted, and in August they started the return. I sent the hardware back, they confirmed on 22 August that they had it, and UPS shows it delivered to their repair centre on 26 August. I have heard nothing at all since. No refund, no reply, three weeks of silence.

I want to be fair about how this started, because the first round was handled well. The board died in late January, I sent photos and video, support looked at it and replaced the board within a week. That replacement arrived, I fitted it, it worked, and I sent the old one back. If it had ended there I would have had nothing but good things to say.

Here is roughly how the rest went.

  • 21 March. The second board dies exactly the same way. No charging light, no power.
  • March and April. Weeks of diagnostics. Mainboard resets, the onboard power switch with a spudger, different chargers including one from a Dell, everything stripped out down to a bare board, photos of the board with the stickers peeled off, video of the LEDs. Nothing.
  • 10 April. Framework takes the whole laptop back to the repair centre.
  • 2 May. They tell me the repaired laptop has shipped.
  • 4 May. Support emails me attaching what they call the repair report, and tell me a component was damaged by overheating, and that this happens when airflow is reduced, “typically from dust buildup or limited ventilation over time.” That board went in on 13 February and was dead on 21 March. Thirty-six days. There was no dust in it. I have since read the attachment properly. It is from LetMeRepair in Bautzen, it says the mainboard was replaced, and it gives no cause at all. No overheating, no dust, no ventilation, nothing.
  • 18 May. Sixteen days after being told it shipped, I still have nothing, so I ask. Only then do I find out UPS could not deliver it, held it at an access point, and sent it back because the holding period expired. Nobody had told me a delivery was attempted.
  • 19 May. I ask why two boards have now died. Support tells me they are “unable to determine the exact cause” and that their repair centres “do not perform board-level repairs and only replace parts/components.”
  • 6 June. After a lot of chasing, Framework offers a complete replacement laptop free of charge. Genuinely a good gesture and I said so at the time.
  • 12 June. It arrives.
  • 17 June. I assemble it. No display. Blink code says DDR not initialised.
  • 18 June. After a round of photos, support tells me they can see a small round metal object inside RAM Slot 1 and asks me to try to move it with the Framework screwdriver. I do what they ask.
  • 19 June. I write back the next morning to say I think I have made it worse. That gets escalated, and nobody ever comments on the damage.
  • 22 June. Another board is sent.
  • 29 June. It is dead when I install it. The packaging had already been opened before it reached me. I photographed that and raised it. It was never answered.
  • 2 and 3 July. I run the full memory isolation test they ask for, every module in every slot, singly and paired, plus an external monitor. With one module the board completes POST with no error code at all and there is still no display. If the memory trains and posts cleanly and the machine is dead, the board is the problem. Nobody responded to that point.
  • 4 July. I send a formal escalation.
  • 15 July. Support offers me a refund or an engineering investigation. I reply the next day saying I would rather have the engineering route if it might actually end in a working machine, and I ask four questions before committing.
  • 16 July to 18 August. Silence. Thirty-three days. I chase on 22 July and again on 29 July. On 3 August I give up on the engineering option and ask for the refund instead.
  • 18 August. Support replies for the first time in over a month, says they can “make an exception and allow the return”, and starts the process. The wording is that they will “initiate a refund if everything is in order”.
  • 21 August. I hand the board to UPS.
  • 22 August. Support confirms they have it and says the refund will follow.
  • 26 August. UPS shows it delivered and signed for at the repair centre.
  • 10 September. I send back the last remaining mainboard too, under a second label.
  • Today. Nothing. No refund, and no contact of any kind since 22 August.

The part I keep coming back to is the 4 May email. I was told the failure was caused by dust and poor ventilation, which puts it on me, and the repair report was attached as backing. The report does not say that. It says the mainboard was replaced and nothing more. Two weeks later I was told they cannot determine the cause and do not do board-level diagnosis at all. So the explanation that blames me was never supported by the document sent to support it, and it is still sitting on my file as the official reason a board that had run for 36 days died.

I am not naming any of the support staff involved. Individually they were mostly polite and several of them clearly tried. This is not about any one person.

Where I am now is the worst of both. Framework has had my money since November and my hardware since August, and I have neither. I work as a software engineer and I have spent six months on an old slow machine because of this.

One thing I should say for completeness, because I only worked it out this week. The return label I was sent on 18 August was sized for 3 kg and 46 x 28 x 8 cm, and the email referred to “your order” and “your items” without listing anything. I read it as being for the failed mainboard and sent that back on its own. I now think it was meant to cover the whole order. I have asked support for a second label and I will return the display, top cover, battery, webcam, hinges, speakers, Wi-Fi card, input cover and bezel the day it arrives. If an incomplete return is why the refund has stalled, nobody has mentioned it to me in the three weeks since.

What I am asking for is simple. Someone with authority to look at the case, pay the refund I was offered in July and accepted in August, and tell me why five boards failed.

Happy to DM the order number, case number and RMA to any staff member who wants them, along with the full correspondence and the photos and videos from every stage.

For a company that was founded with repairability and upgradeabilty at its core this has been a shocking experience.

Has anyone else had a run of Ryzen AI 300 boards fail like this? I would like to know whether I have been very unlucky or whether there is a pattern.

2 Likes

Sorry to hear of your trouble!

I don’t think there is a customer who can give a solid overview on this; nearly every customer, by definition, has such a small perspective on the overall support operation that their good or bad opinion isn’t likely to be statistically significant.

However I do think it is good to post here. We have a few customer-facing employees here, and they are exceptionally helpful. They are a bit maxed out, what with the current phased product launches, but I presently would be confident they can poke the right department.

Hi Thomas, first thanks for bringing this to my attention. I’m forwarding this to the team to look into. I want to apologize for the delay. I can see you were escalated, but there has been an increase in support tickets recently that added to the delay in response. That added with further responses may have reset your queue time as mentioned in the response post. Either way, I apologize for this long lead time you’ve experienced and I’ve forwarded this to the team.

1 Like

Hi. I am just another user.
I think it is a real shame that FW don’t do board level repairs or at least board level investigations in situations like this.
If FW did, they would discover hardware problems and hardware weaknesses and then design fixes in the next release.
Outcome, fewer returns in future == more profit.
An example from above. It was unproven, but FW suggested a component failed due to overheating as a result of lack of airflow.
FW should design hardware fixes for that.
Solution here is to add extra temp sensors on the heat sink, near the fan, so that it can detect if the air flow cooling not actually cooling properly, send a notification to the user, and protect itself by either throttling or switching off itself.

If a user saw a warning message like this, they could do something about it. Maybe discover a usage pattern that causes the cooling to fail. There have been examples of laptops waking up from sleep in backpacks, and overheating as a result. Another possible cause might be blocking of fan inlet, outlets by placing the laptop on a soft surface like a bed or sofa, but without the detection and warning message, the user has no idea they should move the laptop to get better air flow.