Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.
You are now subscribed
Your newsletter sign-up was successful
An account already exists for this email address, please log in.
Subscribe to our newsletter
Micron has finally joined the 3GB GDDR7 party and has officially introduced 3GB modules to compete with Samsung and SK Hynix's 3GB variants. The memory maker officially announced its larger GDDR7 version in a blog post, revealing that its new 3GB ICs operate at 36 Gbps.
The new modules are 12.5% faster than the first GDDR7 modules that hit the market, which achieved 32 Gbps of bandwidth. However, Micron's new modules are noticeably slower than Samsung's competing solutions, which can run at up to 42.5 Gbps. SK Hynix similarly has outgoing GDDR7 modules that can scale up to 40 Gbps, and is working on future models that can reach 48 Gbps.
However, in Micron's favor, there are no outgoing Nvidia graphics cards that take advantage of GDDR7 speeds anywhere near 40 Gbps or faster. The highest GDDR7 transfer speeds on outgoing GDDR7 GPUs include the RTX 5080, which has its memory modules running at 30 Gbps. Other cards, such as the RTX 5090, use even lower memory speeds of 28 Gbps. That said, we have seen memory overclocks on the RTX 5070 Ti and RTX 5090D of up to 34 Gbps on SK Hynix and Samsung GDDR7.
Article continues below
As a result, there's not much reason for Micron to introduce new GDDR7 memory modules with blisteringly fast speeds of 40 Gbps or newer. The next generation RTX 60 series or next generation AMD Radeon GPUs might take advantage of higher memory speeds, but it's not required right now on existing GDDR7 GPUs.
Beyond speeds, having a third manufacturer for 3GB GDDR7 will be highly beneficial for Nvidia and other GPU makers. Nvidia is already taking 3GB GDDR7 chips from both Samsung and SK Hynix, and likely will start introducing Micron 3GB GDDR7 to help compensate for the outgoing NAND flash/memory shortage. Currently, the only Nvidia GPUs that use 3GB GDDR7 are the laptop-variant of the RTX 5090 and the RTX Pro 6000 Blackwell workstation GPU. If Nvidia ever decides to launch the RTX 50 Super series, it is expected that all rumored GPUs in that lineup, the RTX 5080 Super, 5070 Super, and 5060 Super, will use 3GB modules.
Micron can respectfully eat poo. Made an account just to say that. Lurked forever, but recent moves by so many tech and software companies have been really bad and I guess today's the day that I sign up to drop unhelpful venting commentary in the comments. Micron pulling Crucial out of the consumer market is just the latest straw. I wish them the worst in whatever they do going forwards.
Edit: it won't be ALL venting obv but today it is for sure 🙃
recent moves by so many tech and software companies
Okay, but you understand that this particular move makes it more likely that we get a "Super" refresh of RTX 5000 GPUs, right?
It just seems like an odd point to trigger such a response, but your overall sentiment is not uncommon around here.
kaiyulie said:
Micron pulling Crucial out of the consumer market is just the latest straw.
I think the main effect of that is bad publicity, rather than making any real difference in market availability of SSDs or DIMMs. Other retail SSD and DIMM brands can & do still source chips from Micron.
Had Crucial not been furloughed, they'd have had no better pricing or availability than the rest of the brands out there. Just like all the other memory makers, Micron isn't going to sell anything for below market value.
I think it's all very well for them to come out with these new modules but how much good does it do for the consumer in the end if their production capacity is heading more and more towards HBM?
Nvidia is already taking 3GB GDDR7 chips from both Samsung and SK Hynix, and likely will start introducing Micron 3GB GDDR7 to help compensate for the outgoing NAND flash/memory shortage.
