Originally Posted by agentkay
Better than the vaporware that Havok/ATI promised all the years. Using PhysX and CUDA to close the gap until proper DX11/Compute Shader hardware by Nvidia is released, is a solid strategy. This approach beats the useless DX10.1 gimmick anytime.
Whats even cooler when Havok starts to appear in games for real physics, Nvidia will be able to run it as well. They have already shown working OCL drivers for CUDA using a Havok engine, so AMD/ATI wont have a leg up there, as Havok will also be running on Nv hardware aswell.