https://wccftech.com/amd-radeon-vii-worlds-first-gaming-7nm-graphics-card-review-roundup/

Today is the day and now is the time, AMD Radeon has unleashed the worlds first 7nm gaming graphics card, the Radeon VII. A trimmed down but faster version of the last generation Vega architecture with some reworking on the 7nm node along with double the memory should prove this to be a quite the contender in the high end space.

The heart and soul of the aptly name Radeon VII is the 7nm variant of the Vega architecture dubbed Vega 20.  Rather than the 64 or 56 CU variants that were used in the Radeon RX Vega 64 and RX Vega 56 AMD decided to go with a 60 Compute Unit variant.  From numbers run in the past a RX Vega 56 that was vBIOS modded to run at RX Vega 64 clock speeds on the Core and HBM2 found that it was so close in performance that it made one question the need for the 64 Compute Unit model in gaming.  So the decision to go with lower than the maximum Compute Unit design makes sense here since AMD has committed the space and power saving features of the 7nm node to put that budget towards clock speed.  Because of that the Radeon VII is targeting a 1800MHz clock rate which is much higher than what any of the previous RX Vega cards would achieve on air.  The only one that came close to those clock speeds was the RX Vega 64 Liquid Cooled Edition that, while came with a nice watercooling design, carried a much higher TDP and power draw to reach those sustained clocks.  The Radeon VII aims to break that barrier all on air and this time packing 16GB of HBM2 with a mind melting 1TB/s of memory bandwidth, which should help give it an edge at higher resolutions where memory tends to be a constraint.