With the launch of NVIDIA's RTX 50 series GPUs, as is tradition, we were introduced to a brand-new design for their signature "Founders Edition" cards. While brands like ASUS, Gigabyte, MSI and others will release their own designs a little while after the official launch, NVIDIA always releases their own take on reigning in the horsepower of their latest creation first in the form of a reference card.

Historically, NVIDIA hasn't been the gold standard for GPU cooling and power delivery. The Founders Edition cards were more for consumers who wanted to get their new GPU as soon as possible rather than the best possible implementation, but that might change with this product cycle. NVIDIA's new cooler design for the RTX 50 series looks to be extremely well engineered. We don't have official benchmarks yet, but if the teardowns are anything to go by, they might have something special on their hands.

Founders Edition cards through the generations

They haven't always been good

NVIDIA has often been unable to match the designs coming from their board partners, both in terms of thermal performance and robust power delivery. This was especially true in the era of blower style coolers as seen on the 9 and 10 series GTX cards. These cards didn't have unreasonable temperature levels out of the box, but that style just wasn't able to get the most out of the card compared to an AIB open-air design. Blower style coolers are limited by the single fan, along with the size and efficiency of any vapor chamber and heatsink that is used with the card.

The RTX 20 series was a huge departure from the norm for NVIDIA's reference cards. Instead of one single radial fan in a blower-style arrangement, they opted for two axial cooling fans that you would more commonly see on AIB designs. Besides being put together with a fair bit glue, the coolers were a welcomed change by most and performed better. Most third-party vendors still had an edge though, with cards that were able to push the limits further than NVIDIA's reference cards could.

The 30 and 40 series RTX reference cards further innovated with the axial fan design, but especially in the case of the 4090, the cards kept ballooning in size as the thermal load continued to increase. What was once a comfortable two-slot card turned into a pretty massive three-slot monster. NVIDIA used the space on the cooler well, opting to essentially cover the thing in heatsinks to radiate heat way from the GPU and PCB.

👁 nvidia geforce rtx 4070 super founders edition stood up on table to show pcie tab
How much is too much to spend on a gaming GPU?

What is realistically the right price for a gaming GPU to enjoy a reasonable gaming experience?

Impressive feats of engineering

NVIDIA needed to find another solution for the 50 series

Credit: Source: NVIDIA

The RTX 5090 Founders Edition, despite being very thermally demanding with a TDP of 575w, has been brought back to a 2 slot design, impressively enough. They've designed a dual-axial blow through design with a vapor chamber in the middle. The PCBs of these cards continue to get smaller, and this one is the smallest we've seen for a flagship card, with a very dense, square shape. Underneath the vapor chamber is the GPU, but instead of normal thermal interface material like a paste, NVIDIA has opted for liquid metal to further increase the thermal performance of the card.

The main concern with putting liquid metal on a GPU like this is the risk of it leaking out and causing damage to the card, but NVIDIA has designed a rubber gasket around the GPU die to prevent this from happening. Along with the rest of the changes, this is some impressive engineering from NVIDIA. We obviously haven't seen definitive numbers on how this performs yet, but all signs point to this cooling solution being a very strong one.

👁 An image showing Intel and ASRock's immersion tank PC that kept on a desk.
5 unique ways to cool your PC without liquid cooling

AIOs and custom watercooling loops might be cool, but why don't you try something that stands out?

Are NVIDIA's board partners in trouble?

The Founders Edition could be the card to beat

We've never truly seen a reference card be the king of a product stack. AIBs can usually offer slightly more in terms of performance and features, but NVIDIA looks to make things even more difficult for their partners with their 50 series design. Don't get me wrong, it's not a bad thing that the Founders Edition cards have the potential to be really good, but you have to hope that there are still players in the GPU space that want to push limits. NVIDIA being the sole manufacturer of their own GPUs might not be the best thing for consumers.

Historically, partners EVGA would've been keen to push the boundaries of a GPU like this, but their partnership with NVIDIA dissolved many moons ago. It would've been amazing to see what they could've done with a PCB design like this one. Current contenders could be companies like GALAX with their Hall Of Fame line of cards, or even PNY, who have been rumored to be interested in a partnership for next-gen cards with overclocking king and former EVGA ambassador Vince "K|NGP|N" Lucido. It's yet to be seen, but along with the usual suspects of ASUS, Gigabyte, MSI, and ZOTAC, everyone might have to step up their game further to offer compelling value over the reference cards.

👁 Render of the RTX 5070 from Nvidia
I'm worried that GeForce RTX Founder Edition cards won't get waterblocks

I've been into custom watercooling my PC for a long time now, long enough to remember when all we had was aquarium parts and universal blocks that weren't pretty but got the job done. But I'm worried that the Founder's Edition GPUs won't get waterblocks because of how complicated they are with the PCIe and I/O panel on add-in cards. That might mean AIB cards are the way to go for watercooling, and I don't like paying the extra for factory overclocks or overbuilt coolers when I'm planning on watercooling the card. Are you planning on a custom watercool build or do you prefer the simplicity of air cooling?