DeepSeek Unveils V4 AI Models Amid Huawei Tie‑Up and US IP Accusations
DeepSeek unveils its V4 AI models, Pro and Flash, with Huawei's full support for Ascend systems. This launch follows US allegations of industrial-scale IP theft from AI labs.

DeepSeek launched its fourth-generation V4 artificial intelligence models, including Pro and Flash versions, following an announcement of full support from Huawei's Ascend supernode systems. This development occurred just one day after the White House accused China of industrial-scale intellectual property theft from US AI laboratories.
DeepSeek, a notable Chinese artificial intelligence startup, has introduced its V4 AI model series, replacing the previous V3.2. The company gained attention for delivering strong AI model performance at lower operational costs compared to some US rivals. The new V4 series offers two distinct versions: V4 Pro, designed for enterprises and advanced users seeking robust reasoning capabilities and broader world-knowledge performance, and V4 Flash, positioned as a faster, lower-cost option for wider deployment. Both models now support an ultra-long context window of one million tokens, allowing significantly more information to be processed in a single session.
The V4 launch included a significant announcement from Huawei, stating its Ascend supernode systems will fully support DeepSeek's new V4 model series. These supernode systems provide the high-performance computing infrastructure essential for training and deploying advanced artificial intelligence models. This partnership signifies a deepening cooperation within China's domestic technology ecosystem. This unveiling followed closely on the heels of the White House issuing accusations of industrial-scale intellectual property theft from US AI laboratories by China, made just one day prior to DeepSeek's V4 presentation.
Huawei's explicit commitment to supporting DeepSeek's V4 models strengthens China's strategic push for a self-sufficient artificial intelligence supply chain. This move aims to bolster domestic chip and infrastructure development, reducing reliance on foreign semiconductor technology following US export restrictions. The timing of DeepSeek's V4 launch, directly following US accusations of IP theft, places these new models within a heightened geopolitical context. This situation highlights ongoing tensions between the US and China over AI leadership, technological innovation, and data security. The market will now closely monitor DeepSeek’s performance claims against industry benchmarks and the broader adoption of V4 models within the evolving global AI landscape. Future developments in US-China tech policy will also be key to watch.
Continue reading
More in this thread
Micron’s Profit Margin Surges to 57.8% on AI‑Fueled Revenue Triple‑Jump, While AWS Posts Fastest Growth in 13 Quarters
Alex Mercer
Maryland's 45,000 aerospace jobs at risk as experts warn space debris could trigger GPS‑destroying chain reaction
Alex Mercer
Nvidia’s $1 Trillion AI Chip Orders and Broadcom’s $100B Forecast Underscore Nasdaq AI Upside While Microsoft Remains 20% Below Its Peak
Alex Mercer
Conversation
Reader notes
Loading comments...