Listen to this Post
Introduction:
Chinese AI startup DeepSeek has encountered significant setbacks in the development of its R2 AI model, originally slated for release in May 2025. The company faced challenges when attempting to train the model using Huawei’s Ascend chips, leading to performance issues that hindered progress. Despite efforts to resolve these problems, DeepSeek ultimately reverted to using Nvidia chips for training and Huawei’s for inference. This delay underscores the ongoing limitations of China’s semiconductor technology and the challenges faced by domestic companies in achieving technological self-sufficiency.(implicator.ai, Financial Times)
Summary:
DeepSeek, a Chinese artificial intelligence company, has postponed the launch of its R2 AI model due to difficulties encountered when training the model on Huawei’s Ascend chips. The company had initially planned to release the model in May 2025 but faced persistent technical issues that prevented successful training runs. In response, DeepSeek reverted to using Nvidia chips for training, while relying on Huawei’s Ascend chips for inference tasks. Huawei engineers were dispatched to assist in resolving the issues, but as of now, the model has not completed a successful training run on the Ascend chips. This situation highlights the ongoing challenges faced by Chinese tech companies in developing advanced AI models using domestic semiconductor technologies. Despite the setbacks, DeepSeek remains committed to advancing its AI capabilities and is exploring alternative solutions to overcome the current limitations.(Financial Times)
What Undercode Says:
The delay in DeepSeek’s R2 AI model launch due to issues with Huawei’s Ascend chips highlights a critical juncture in China’s pursuit of technological independence. While the Chinese government has been actively promoting the use of domestic technologies to reduce reliance on foreign suppliers, this incident underscores the practical challenges faced by companies in achieving this goal. The performance limitations of Huawei’s Ascend chips, particularly in training large-scale AI models, are evident. DeepSeek’s decision to revert to Nvidia chips for training, despite the geopolitical tensions surrounding chip exports, reflects the current gap in performance between domestic and foreign semiconductor technologies.(Financial Times, Reuters, TrendForce)
This situation also raises questions about the scalability and long-term viability of China’s push for self-reliance in AI development. While strides have been made in areas such as software optimization and model efficiency, the underlying hardware still poses significant challenges. The hybrid approach adopted by DeepSeek—utilizing Nvidia chips for training and Huawei’s for inference—may be a temporary solution, but it highlights the need for substantial advancements in domestic chip technology to support the next generation of AI models.
Furthermore, this development could have broader implications for the global AI landscape. As Chinese companies strive to develop competitive AI models, the interplay between domestic and international technologies will shape the future of AI innovation. The outcome of DeepSeek’s efforts could influence the direction of AI development not only in China but also globally, as companies and governments navigate the complexities of technological sovereignty and collaboration.
Fact Checker Results:
DeepSeek’s R2 AI model launch has been delayed due to technical issues with Huawei’s Ascend chips, as reported by the Financial Times and Reuters. (Financial Times)
Despite assistance from Huawei engineers, DeepSeek has not yet completed a successful training run on the Ascend chips. (Financial Times)
The company has reverted to using Nvidia chips for training and Huawei’s for inference tasks, highlighting the performance limitations of domestic semiconductor technologies. (Financial Times)
Prediction:
Given the current challenges faced by DeepSeek in training its R2 AI model using Huawei’s Ascend chips, it is likely that the company will continue to rely on Nvidia chips for training purposes in the near future. This reliance may persist until significant advancements are made in the performance and compatibility of domestic semiconductor technologies. Additionally, the hybrid approach adopted by DeepSeek could become a model for other Chinese AI companies facing similar limitations. However, the ongoing issues underscore the need for substantial investment in research and development to bridge the technological gap and achieve true self-reliance in AI hardware.([Wikipedia][6])
[6]: https://en.wikipedia.org/wiki/DeepSeek?utm_source=chatgpt.com DeepSeek
🕵️📝✔️Let’s dive deep and fact‑check.
References:
Reported By: xtechnikkeicom_079bc20707117cb00ca0292a
Extra Source Hub:
https://www.quora.com/topic/Technology
Wikipedia
OpenAi & Undercode AI
Image Source:
Unsplash
Undercode AI DI v2
🔐JOIN OUR CYBER WORLD [ CVE News • HackMonitor • UndercodeNews ]
📢 Follow UndercodeNews & Stay Tuned:
𝕏 formerly Twitter 🐦 | @ Threads | 🔗 Linkedin | 🦋BlueSky | 🐘Mastodon
