DeepSeek V4 AI model release is reportedly imminent, according to a Financial Times report published Friday (Feb. 27), marking a fresh moment in China’s effort to compete with leading American AI developers.

The FT, citing two sources familiar with the matter, reported that DeepSeek plans to unveil V4, a new “multimodal” large language model capable of generating pictures, video and text. The timing of the release could position the company prominently ahead of China’s annual parliamentary “Two Sessions” meetings, which begin March 4.

DeepSeek V4 AI Model Release and Multimodal Ambitions

According to the Financial Times, the DeepSeek V4 AI model release would represent the company’s first major launch since its R1 reasoning model debuted in January last year.

At that time, DeepSeek said R1 could perform at the level of top American AI models while using just a fraction of the computing power. That claim reverberated across the U.S. technology sector, with some experts comparing the moment to the Soviet Union’s Sputnik launch in 1957.

The FT report also noted that DeepSeek worked with domestic chipmakers to prepare the new model. Specifically, it said the company collaborated within the DeepSeek China AI Huawei Cambricon chips ecosystem to optimize V4 for Huawei and Cambricon’s latest hardware releases.

By aligning closely with Chinese semiconductor firms, DeepSeek appears to be reinforcing a domestic supply chain strategy as geopolitical and trade pressures continue to shape the global AI landscape.

The launch’s proximity to the “Two Sessions” meetings could further elevate DeepSeek’s status as a national AI champion, according to the FT.

China US AI competition DeepSeek vs OpenAI Dynamics

The DeepSeek V4 AI model release arrives amid intensifying China-US AI competition DeepSeek vs OpenAI narratives. DeepSeek’s earlier R1 announcement prompted analysts to reassess assumptions about cost structures and efficiency in AI model development.

Gokul Naidu, a consultant for SAP, told PYMNTS at the time that DeepSeek could catalyze a shift in Silicon Valley’s approach to AI. He said the company’s dramatically lower cost structure was forcing businesses and developers to reconsider efficiency-driven growth rather than prioritizing expansion at any cost.

Since the R1 debut, however, the Financial Times reported that DeepSeek has mainly issued incremental updates instead of entirely new models. That pause has allowed Chinese rivals, including Alibaba, to capture some demand for lower-cost, open-source AI systems.

The upcoming release of V4 could therefore serve as a renewed test of how effectively China’s AI sector can narrow the competitive gap with American firms such as OpenAI.

In related AI commentary, industry pioneer Andrew Ng said last week that artificial general intelligence remains decades away. Speaking to Fast Company, Ng defined AGI as AI capable of performing the same intellectual tasks as humans, such as driving a truck or writing a Ph.D. thesis.

“We’re still very far from AI meeting that definition of AGI,” said Ng, founder of DeepLearning.AI and Coursera, executive chairman of LandingAI and founding lead of Google Brain.

He added that while some companies use looser definitions of AGI, under his interpretation, the industry remains “very, very far away” from machines matching human intelligence.

As anticipation builds, the DeepSeek V4 AI model release is shaping up as more than just a product update. It represents another milestone in the evolving global race to define the next generation of AI capabilities.