DeepSeek’s New Model Receives Tepid Response
A little over a year ago, the Chinese artificial intelligence lab DeepSeek made headlines by releasing models that rivaled the best in the West at a fraction of the cost. This breakthrough moment briefly unsettled the market value of major AI infrastructure providers like Nvidia. However, the recent launch of DeepSeek’s new model, v4, has not generated the same excitement, raising questions about its impact and the current state of AI competition.
DeepSeek’s v4 Model Performance
DeepSeek’s latest model, v4, has demonstrated impressive capabilities, closely matching the performance of leading American AI models released just months prior. According to the company’s tests, the most powerful version of v4, known as the “Pro” system, falls slightly short of its Western counterparts. This new model is also positioned as a cost-effective option for customers, with an introductory price that is a mere one-thousandth of the cost of top American models for certain applications. Even after the promotional pricing ends on May 7, v4 will still be significantly cheaper, costing between 10% and 25% of its American equivalents. Despite these advantages, the excitement surrounding its release has been muted compared to the initial response to DeepSeek’s earlier models.
Challenges in Development and Competition
While DeepSeek’s v4 model showcases strong performance, it appears that its development was not as cost-effective as its predecessor. In 2025, the lab reported that training its AI cost around $6 million, a figure that was notably lower than Western standards. However, the technical white paper for v4 does not provide any cost estimates for its training, suggesting that the development process may have required substantial resources. The 16-month gap between the releases of v4 and its predecessor indicates that significant processing power was likely utilized in its training. Furthermore, DeepSeek faces increasing competition from other independent labs and major Chinese tech companies, such as Alibaba and ByteDance, which have been making strides in the AI landscape.
Shifting Focus in China’s AI Landscape
In recent months, the focus in China’s AI sector has shifted towards applications built on AI technology rather than just the models themselves. Companies like Alibaba have integrated their Qwen model into various aspects of their business, providing a “digital workforce” for merchants on their e-commerce platform. The race among China’s internet giants is now centered on developing AI-powered “super apps” that can streamline a variety of digital transactions. This shift indicates that merely having advanced AI models is no longer seen as a viable path to monetization in the competitive market.
Government Influence and Future Prospects
DeepSeek’s operations are also affected by increased government oversight. The Chinese government has been promoting the use of domestically produced chips, particularly those made by Huawei. Reports suggest that DeepSeek initially attempted to train its new model using Huawei’s chips but ultimately reverted to Nvidia’s technology, which added both cost and time to the development process. The government’s recent decision to block the acquisition of Manus, another prominent AI company, by Meta underscores the ongoing challenges local AI firms face in navigating state regulations. As the competition intensifies, the future of DeepSeek and its v4 model remains uncertain, especially as the landscape evolves and the balance of power shifts in the AI industry.
Observer Voice is the one stop site for National, International news, Sports, Editor’s Choice, Art/culture contents, Quotes and much more. We also cover historical contents. Historical contents includes World History, Indian History, and what happened today. The website also covers Entertainment across the India and World.
Follow Us on Twitter, Instagram, Facebook, & LinkedIn
