Recent generations of flagship graphics cards tend to have increased maximum power consumption. The RTX 4090 can pull out more than 450W, and the RTX 3090 Ti can do the same, while AMD's RX 7900 XTX is rated at 355W. Factory overclocked models usually cost a little more. But these are rookie numbers. Nvidia's next-generation GB200 AI accelerator can reportedly pull up to 1,000W.
The source of the 1,000W claim is a statement made by Dell COO Jeffrey Clarke during Dell's Q4 2024 earnings call (via The Register). This 1,000W claim of his is more than just a rumor, as he is likely more informed than anyone else.
Clarke's actual words are, “Our characterization on the thermal side is that we don't really need direct liquid cooling to reach an energy density of 1,000 watts per GPU, and that's what we'll get with next year's B200. “is.
Note that although he mentions the B200, the B200 isn't actually on Nvidia's roadmap. The next generation flagship enterprise GPU is called B100. Based on Blackwell architecture. However, there is also his GB100, which combines a B100 GPU with a Grace CPU. Perhaps Clark simply misspoke and was referring to the latter.
Data center GPUs are very different from consumer gaming GPUs, where things like fan speed, physical size, and power consumption are secondary considerations compared to outright performance. The GB200 is all about performance. Various variations of Blackwell-based accelerators will ultimately form the backbone of cloud AI systems in 2025 and beyond.
We all know by now how big business AI is. Nvidia is now the world's third-largest company by market capitalization, recently overtaking Saudi-based oil giant Aramco, with only Apple and Microsoft remaining above it. And it's rapidly gaining popularity, showing that the demand for high-performance AI accelerators remains insatiable.
I can't imagine ever seeing a 1,000W desktop graphics card. This is quite an excessive step for a typical ATX PC, but with 450W becoming the new baseline for high-end gaming GPUs, we expect things to step up again. The 500W wall has already all but disappeared, and if Nvidia chooses to release his RTX 4090 Ti, we can imagine 600W or more becoming a reality. Take a look at the leaked images of the giant quad-slot Founders Edition cooler.
Personally, I hate the idea of ​​cards using that much power. I use his 450W RTX 4090 regularly and it gets pretty hot in my office room during long sessions. He doesn't want to think about what will happen to cards over 600W during the warmer months of the year.
Whatever final form the future RTX 5090 takes, it's going to be a hot, power-hungry card. If a high-end Blackwell-based GPU has a TDP above 500W, we can expect quad-slot or AIO cooling cards to become the norm, especially overclocked models.
We hope cards like the RTX 5060 and RTX 5070 can maintain the great power efficiency and performance-per-watt characteristics of cards like the RTX 4060 and the recently released RTX 4070 Super.