NVIDIA RTX 50 coolers under active development

With great power comes a great electricity bill.

Prototype RTX 5090/TITAN RTX ADA cooler vs. TITAN RTX (Ampere)

Since the RTX 40 series dropped, there’s been talk about NVIDIA experimenting with a triple-fan and quad-slot setup for a new GPU. People thought it was just a rumor, especially after the RTX 4090 hit the scene. However, later photos have surfaced proving that this wild cooler concept is real. It’s got a unique perpendicular PCB layout and a huge quad-slot cooler with a third fan hidden inside the heatsink. Early whispers suggested it could handle up to a 800W total graphics power (TGP).

Now, according to Benchlife, NVIDIA’s busy testing coolers that can handle up to 600W. That means the next-gen GPUs from the Blackwell series might stick with the power specs we’ve got now with Ada. The Founders Edition model based on the RTX 4090 SKU already supports 600W, maxing out what you can tweak in the BIOS.

According to information from a familiar cooling module manufacturer, NVIDIA has already conducted relevant tests and verification on cooling modules for the Blackwell GPU architecture’s GeForce RTX 50 series graphics cards, evidently preparing for the launch of the GeForce RTX 50. Although there’s no specific timeline, currently, about 4 plans are underway, with power ranging from 250W to 600W.

— Benchlife

It’s no surprise that NVIDIA is always inventing new coolers. They’ve likely been at it for months, maybe even years, gearing up for whatever’s next on the horizon. For cooling engineers, the only thing that matters is how large the GPU is and how much power it will require, the rest is up to designers and marketing who will decide on the final GPU specs.

A very interesting video on a subject was shared by Gamers Nexus in 2022, whose guest was NVIDIA Thermal Engineer, definitely worth a watch if you missed it. It discusses the process of choosing the right cooler, the validation phase and explains what vapor chambers do.

The Ada Lovelace GPU lineup in the current generation ranges from a TDP of 115W to 450W on desktops, with the maximum TGP stretching from 130W to 600W. You can verify these figures using GPU-Z or by checking the TechPowerUp GPU BIOS database for each model.

As you can see, just because we hear rumors about 600W coolers, it doesn’t mean all board partners will use maximum available power within the BIOS. The actual GPU may and probably consume close to what is the current TDP of RTX 4090. Furthermore, if NVIDIA is sticking to its 12VHWPR power standard, these connectors will not be able to deliver more than 600W.

Source: Benchlife via Wccftech