DeepSeek's R2 Model Launch Postponed

In a surprising turn of events, DeepSeek, a leading Chinese artificial intelligence company, has delayed the launch of its highly anticipated R2 model. According to the Financial Times, the delay stems from technical challenges encountered while attempting to train the model using Huawei's Ascend processors, as part of a broader initiative to reduce reliance on Nvidia's technology.

The Struggle with Huawei's Chips

Despite the initial enthusiasm from Chinese authorities for using domestic technology, DeepSeek faced "persistent" technical issues during the training phase. These complications forced the company to revert to Nvidia chips for training purposes, while still utilizing Huawei's processors for inference tasks. This setback has pushed the launch from its original May timeline.

Huawei's Efforts to Assist

In response to these challenges, Huawei dispatched a team of engineers to DeepSeek's offices in an attempt to facilitate the use of its AI chips for the R2 model's development. However, these efforts did not yield the desired results, as the company was unable to complete a successful training run on the Ascend chip. Despite these hurdles, collaboration between the two companies continues, aiming to ensure the model's compatibility with Huawei's technology for inference.