苹果因M2 Ultra芯片无法满足高级AI需求，转向依赖英伟达

Apple is shifting its reliance to Nvidia GPUs for its AI operations, as its self-developed M2 Ultra chips have failed to meet performance expectations for advanced AI workloads.

Reports indicate that Apple had previously attempted to run Google's Gemini model on its own servers to power the AI capabilities of the new Siri, but the performance of its chips was insufficient for the task, ultimately forcing it to turn to Nvidia GPUs deployed on Google Cloud. Furthermore, the development of Apple's next-generation AI server chip, "Baltra," has experienced delays. The M7 Ultra chip, which is expected to challenge the performance capabilities of Nvidia's Blackwell architecture, is not projected to be released until 2029 at the earliest. (Coin Bureau)