- Deepseek Delays R2 launch due to sustained technical difficulties with Huawei Ascend Chips
- Nvidia H20 systems remain more reliable to AI training than Domestic Ascend Hardware
- Alibaba’s QWEN3 utilizes Deepseeks delays containing core algorithms while improving efficiency and flexibility
The Chinese AI Giant Deepseek has apparently encountered unexpected delays in releasing its latest model, R2, after being exposed to persistent technical difficulties with Huawei’s Ascend Chips.
The company had been invited by Chinese authorities to adopt domestic processors instead of relying on Nvidia’s H20 systems, which are generally considered more mature and reliable.
Despite the fact that Huawei engineers were on site to help, Deepseek could not complete a successful training with the help of Ascend Chips -and as a result, the company depended on Nvidia hardware for training while using Ascend for inference tasks.
Technical challenges delay R2 development
The R2 launch, originally scheduled for May 2025, was postponed due to these technical obstacles and longer than expected data marking for the updated training data set.
Deepseek founder Liang Wenfeng allegedly expressed dissatisfaction with the model’s progress and emphasized the need for further development time to produce a model capable of maintaining Deepseeks competitive advantage.
Meanwhile, competitors such as Alibaba’s QWEN3 were able to take advantage of this delay as it has incorporated Deepseek’s core education algorithms while improving efficiency and flexibility, showing how fast AI ecosystems can develop even when a single start -up is fighting.
Peking’s wider push for AI-self supply has put pressure on domestic companies to adopt local hardware.
In practice, however, this strategy has revealed gaps in stability, inter-chip connection and software ripening between Huawei chips and NVIDIA products.
Developers continue to play a crucial role in designing the success of AI -Ecosystems -Nvidia has emphasized maintaining access to Chinese developers is strategically important, and warning that the limitation of technology -admission can harm economic and national security interests.
Chinese AI companies, meanwhile, have to balance government pressure with practical realities in the development and implementation of LLMs.
Despite these setbacks, Deepseeks R2 model can still be released in the coming weeks.
The model is likely to face control of its performance in relation to rivals trained on more mature hardware, giving a clear example of the tension between political ambitions, technical capacity and AI implementation in the real world.
Via Arstechnica
