苹果因M2 Ultra芯片无法满足高级AI需求，转向依赖英伟达

Odaily Planet Daily News Apple is turning to NVIDIA GPUs to support its AI business, as its custom-designed M2 Ultra chip failed to meet expectations when running advanced AI workloads.

Sources indicate that Apple had previously attempted to use its own servers to run Google's Gemini model, providing AI capabilities for the new Siri, but the performance of its chips was insufficient, ultimately forcing it to rely on NVIDIA GPUs deployed on Google Cloud. Additionally, development of Apple's next-generation AI server chip, codenamed "Baltra," has faced delays. Meanwhile, the M7 Ultra chip, which is anticipated to rival the performance of NVIDIA's Blackwell architecture, is not expected to launch until 2029 at the earliest. (Coin Bureau)