NVIDIA is currently preparing to launch its new line of graphics cards, this time based on the 5nm Ada Lovelace architecture.
However, we do not yet know which nomenclature NVIDIA will choose, whether it is RTX 40 or RTX 50, as rumors and “leaks” suggest that nothing has been decided yet. (But because the name RTX 50 carries a lot of weight, performance has improved over the current one RTX 30).
But one thing is for sure… They are going to be power consuming graphics cards. So if you still haven’t changed your power supply? Well… Get ready!
NVIDIA tests 900 watt graphics card!
Therefore, it is no longer news to anyone that the new generation of graphics cards will cause a very significant increase in consumption. But at this level? It’s just to keep AMD from escaping, which in turn would have to leverage MCM design to deliver unprecedented performance with its new RDNA3 architecture… NVIDIA is currently forced to turn any extra watts into clean and hard performance!
To try to understand how far it can go with TSMC’s technology produced on 5nm lines, NVIDIA is testing a graphics card based on the AD102 graphics chip (the most powerful in the new series) capable of achieving an incredible 900 watts of TGP power. (Total Graphics Power).
Unlike the current RTX 30 series, where the most powerful GPU (GA102) gives life to more than one card, the AD102 GPU only serves as the basis for the most powerful model, probably the RTX 4090 (or RTX 5090). . However, the RTX 4080 (or 4090) should be based on the more modest and, of course, less hungry AD103 graphics processor.
Still, while the AD103 and AD104 graphics processors are enlivened by graphics cards that are likely to be sold more (RTX 4080/5080 and RTX 4070/5070), which in turn should be far from 900W, the growth is expected to be very significant. energy consumption. So start the math in life.