The article is more or less BS.

For one, the claim is that Intel misleads intentionally. Given the timeline there's no particular reason to think there is something intentional here. The AMD-optimized version of GROMACS just came out ~5 weeks ago.

Second, while this is misleading in a sense, it's in the way that benchmarks are all generally misleading. By their intrinsic nature they don't tell the whole story. (And of course marketing departments cherry-pick benchmarks to tell the story they want to tell. Also by intrinsic nature.)

BTW, this isn't even all that bad for Intel. The AMD chip might outperform the Intel one, but only if your software has been specifically optimized for it.