Western Digital has now tacitly admitted that its SanDisk Extreme Pro portable solid-state drives were critically broken after all — by issuing a firmware update that’ll hopefully keep the 4TB model from wiping your data all by itself. (Seriously.)

But as Ars Technica reports, there’s no mention of any fix for the 2TB model, and I think it’s time to sound the alarm — because this company has been downplaying the issue for months, all while it continues to sell these drives at a steep discount.

Over two months ago, my friend and Verge supervising producer Vjeran Pavic told me he’d lost an entire 4TB SanDisk Extreme Pro worth of video clips. Completely gone with no trace. (Thankfully, he’d already backed up all that footage to the cloud.)

SSD failure does happen, but something weirder was going on here — when he tried adding some of the files back, they started disappearing one by one.

And when he contacted a SanDisk customer support agent, he was told that they couldn’t replace his drive because the 4TB model was being recalled. They worked out a deal to swap it for a pair of 2TB units instead — but he hasn’t actually received them because the company’s RMA system was down amidst a data breach.

Unfortunately, as Ars reports, the 2TB model may be susceptible to the issue as well. Ars Technica editor Lee Hutchinson has had two of the 2TB Extreme Pros die on him, with their entire filesystem wiped, and there are similar complaints online.

I feel guilty for not writing about this issue earlier, but there were reasons we held off. First, there was evidence the 4TB model was still on sale and not actually being recalled. Second, when we contacted WD about the issue in March, the company told us it was “actively investigating” a “small number of reports,” downplaying the problem.

But Vjeran rightly tweeted a warning when he noticed the company had started putting the drives on deep discount:

And today, with WD’s tacit admission, there’s certainly no way that I would buy this drive without independent third-party testing that the issue is gone.