On January 28, Fast Technology reported that NVIDIA's stock price took a significant dive, plummeting 17% at the close of US markets, resulting in an overnight evaporation of USD 588.862 billion (approximately RMB 4.27 trillion) in market value.

Industry observers attribute this downturn to the impact of China's AI large model, DeepSeek, which triggered a collective decline in US chip stocks, with NVIDIA bearing the brunt of the impact.

Despite the sharp drop in its stock price, NVIDIA has commended DeepSeek's R1 model as a remarkable advancement in AI.

According to media reports, an NVIDIA spokesperson remarked, "DeepSeek embodies outstanding progress in AI and serves as an exemplary case of Test Time Scaling. The work of DeepSeek demonstrates how to leverage this technology to develop new models, utilizing widely accessible models while fully adhering to export control regulations in computational terms."

NVIDIA further stated its belief that DeepSeek's breakthrough will stimulate greater demand for its products.

On January 20th of this year, DeepSeek officially launched DeepSeek-R1, which is officially stated to perform on par with OpenAI GPT-4's official version in tasks involving mathematics, coding, and natural language reasoning.

The pricing for the DeepSeek-R1 API service is set at RMB 1 per million input tokens (cache hit), RMB 4 per million input tokens (cache miss), and RMB 16 per million output tokens.