It is well known amongst AI engineers that the total compensation packages at AMD, which include base salary, RSUs (Restricted Stock Units), and bonuses, lag considerably behind competitors like NVIDIA, Tesla Dojo, OpenAI Chip Team, Google TPU, and xAI, etc.
In conversations with top AI software engineers about why they chose not to join AMD, many highlighted that working at AMD software feels like working on porting features that NVIDIA engineers developed two years earlier. In contrast, NVIDIA provides engineers with the opportunity to work on cutting-edge software and to build software for chips used in state-of-the-art models like o3, for both training and inference.
Additionally, engineers attracted to the "David" in a "David vs Goliath" scenario often choose Google TPU or the OpenAI Chip team over AMD. These teams offer significantly better compensation and arguably have a higher probability of success in competing against NVIDIA due to those companies having giant volume of internal workloads to be their own customers. This makes them more appealing options for ambitious engineers.