The world isn't satisfied with a "decent model". The world is trying to reach AGI. Further more, reasoning models require more tokens. The faster the GPU, the

The world isn't satisfied with a "decent model". The world is trying to reach AGI. Further more, reasoning models require more tokens. The faster the GPU, the more thinking it can do in a set amount of time. This means the faster the hardware, the smarter the model output. Again, that reinforces the need for faster hardware. More thinking = smarter models Faster hardware = more thinking Therefore, faster hardware = smarter models