Demand for Huawei's Ascend 950 artificial intelligence chips has surged following the release of DeepSeek's V4 model, which is optimized to run on the Chinese company's hardware. Major Chinese internet firms including ByteDance, Tencent, and Alibaba are actively seeking to secure new chip orders, according to a Reuters report. Cloud computing and GPU rental service providers are also moving quickly to place orders.
The Ascend 950PR chip significantly outperforms Nvidia's H20 chip—the most advanced Nvidia processor previously allowed for sale in China before Beijing blocked its import last year. However, it still trails Nvidia's more advanced H200 chip. Despite regulatory approvals from both the US and China, H200 shipments to China have not begun due to ongoing disagreements over sales conditions, creating an opening for Huawei.
The release of DeepSeek V4 has notably boosted demand for domestic AI hardware, reflecting the impact of US export restrictions. DeepSeek's decision to optimize its model specifically for Huawei hardware signals a shift toward China's homegrown semiconductor ecosystem. Huawei stated that its Ascend SuperNode infrastructure, built on the 950 series chips, will fully support DeepSeek V4 models.
Cloud providers have rapidly deployed the model. Alibaba Cloud's Bailian platform made DeepSeek V4 available on release day, offering V4-Pro and V4-Flash versions at official pricing. Tencent Cloud also launched preview services on its TokenHub platform the same day, deploying the model across domestic infrastructure and its Singapore gateway.
DeepSeek added image and video recognition to its main chatbot, a new feature called "image recognition mode" joining "expert" and "flash" modes. The feature allows the chatbot to understand photos and videos, not just text, bringing it in line with other major AI chatbots.
Remarkably, four Chinese chip companies—Huawei Ascend, Cambricon, Hygon Information, and Moore Threads—confirmed their hardware worked perfectly with V4 from the very first day of release. Huawei's Ascend chips, including the A2, A3, and 950, support both V4-Pro and V4-Flash. This coordinated same-day support across multiple domestic chipsets is a significant milestone previously unattainable outside Nvidia's ecosystem.
The V4 models include two versions: V4-Pro with 1.6 trillion parameters and V4-Flash with 284 billion parameters, both supporting a one-million-token context window. They are released under the MIT open-source license, allowing broad commercial use. DeepSeek is offering developers a 75% discount on V4 until May 5 and indicated that V4-Pro pricing could decline in the second half of 2026 once Huawei's supernodes ship at scale. However, supply constraints are expected to persist until production ramps up.
Huawei plans to ship approximately 750,000 units of the 950PR chip this year, with mass production beginning in April and full-scale shipments expected in the second half of 2026. Output is likely to remain below demand due to US restrictions on advanced chipmaking equipment, limiting China's manufacturing capabilities.
