ByteDance Taps TSMC for In-House AI GPU Production, Reducing Nvidia Dependence
09/20/2024, 01:35 AM UTC
字节跳动携手台积电制造自研AI GPU，减少对英伟达依赖TikTok owner ByteDance taps TSMC to make its own AI GPUs to stop relying on Nvidia — the company has reportedly spent over $2 billion on Nvidia AI GPUs
➀ 字节跳动正与台积电合作开发两款AI GPU，预计于2026年实现量产；➁ 此举旨在减少对英伟达的依赖；➂ 字节跳动今年已花费超过20亿美元购买英伟达GPU。➀ ByteDance is developing two AI GPUs with TSMC for mass production by 2026; ➁ The move aims to reduce reliance on Nvidia; ➂ ByteDance has spent over $2 billion on Nvidia GPUs.
ByteDance, the parent company of TikTok, is reportedly investing heavily in the development of two AI GPUs, set to enter mass production by 2026. These GPUs, which are currently in the design phase, are expected to be manufactured by TSMC. The initiative is a strategic move to reduce dependence on Nvidia for AI hardware, while adhering to U.S. export regulations.
With an investment of over $2 billion in Nvidia AI GPUs this year alone, ByteDance's heavy investment in AI underscores its importance in the company's overall strategy. The development of in-house GPUs is a direct response to the GPU shortages and high prices of Nvidia GPUs, particularly in the context of the U.S. export controls imposed last year.
While ByteDance is unlikely to surpass Nvidia's GPU performance due to U.S. export control restrictions, the in-house GPUs are expected to be significantly cheaper for the company. Additionally, the move requires ByteDance to develop its own software platform and ensure compatibility with its hardware, potentially shifting the landscape of AI hardware and software development in the process.
---
本文由大语言模型（LLM）生成，旨在为读者提供半导体新闻内容的知识扩展（Beta）。
