BlockBeats News, July 24th, Citrini analyst Jukan quoted DeepSeek founder Liang Wenfeng's recent key point, pointing out that AI-driven code generation, along with advanced programming languages like TileLang, is rapidly reducing the barrier to entry into the CUDA ecosystem. Although DeepSeek uses NVIDIA GPUs to train the V3 model, it has significantly reduced its reliance on the NVIDIA software ecosystem through its self-developed compiler and TileLang environment. Previously, Liang Wenfeng predicted that if TileLang and the DeepSeek compiler were ported to Huawei chips, the ecological issue of Chinese chips could be basically resolved within about a year, with the real bottleneck being production capacity. Liang Wenfeng quantified the the gap between Chinese and American chips—hardware efficiency is about 4 times higher, with a time lag of about two years—and revealed that DeepSeek is closely collaborating with Huawei, expecting to receive around 16,000 Huawei AI chips. The Huawei 950 SuperNode can replace the workload of the GB200/GB300. Analyst Jukan qualifies this as "The moat of CUDA is coming to an end" and is very pessimistic about NVIDIA.

Jukan added that this logic is precisely one of the reasons to be bullish on AMD— the progress in AI coding will also naturally accelerate the development of the ROCm ecosystem, helping to narrow the gap with CUDA. AMD recently announced active use of Claude Code for chip design and software engineering when investing in Anthropic. Overall, the progress of AI programming tools is systematically eroding NVIDIA's competitive barriers from the software side. The ecological issue of Chinese chips will be rapidly resolved due to code generation capabilities, AMD will benefit from the acceleration of ROCm, NVIDIA's CUDA moat, which is crucial for developer stickiness, is facing a double squeeze, and catching up in hardware efficiency and production capacity is only a matter of time.