Chips War Heats Up as Google Challenges Nvidia

Google aims to develop its own processing chips, known as Tensor Processing Units (TPUs), in order to reduce reliance on external suppliers like Nvidia, improve the performance of artificial intelligence models, and lower operational costs. Since 2016, Google has been launching generations of TPU chips, which are specifically designed to efficiently handle the mathematical operations necessary for neural networks, whether during training or deployment. Google has crafted its strategy around integrating hardware and software, enhancing performance while reducing energy consumption—an important consideration given the increasing demands of models like Gemini, which require substantial investments in infrastructure and power. Although the company continues to use Nvidia graphics processing units (GPUs), developing its own chips grants greater flexibility, reduces supply chain risks, and is part of a long-term strategy to control the entire AI value chain. This approach aligns with academic and technological trends toward greater customization and independence in the field of artificial intelligence.

articleDetails.fontSize.sizeحجم النص

تنويه: هذا ملخص تم إنشاؤه بواسطة الذكاء الاصطناعي

عرض المقال بالكامل