DeepSeek plans to implement a significant price increase across its AI services, an unusual shift from the disruptive Chinese player that has put pressure on US and domestic rivals.
In a notice posted on August 6, DeepSeek did not specify the exact increases, other than to call them substantial and urge users to plan accordingly. The Hangzhou-based company currently charges US$0.14 (RM0.57) per million input tokens and US$0.28 (RM1.15) per million output tokens – measures of work an AI model does to digest information and produce results – for its V4 Flash, whereas Chinese rival Moonshot prices its Kimi K3 at US$3 (RM12.27) and US$15 (RM61.34) per million of each token type. Anthropic’s state-of-the-art Fable 5 service sets that pricing at US$10 (RM40.89) and US$50 (RM204.46), respectively.
DeepSeek’s strategy has sparked a global debate in recent days over the cost of high-performing artificial intelligence (AI) tools. Its pricing has been so far out of the usual range for AI models that it has prompted the idea of a DeepSeek "death zone,” where costlier or less capable models would be obviated. It’s also raised questions about the initial public offering plans of OpenAI and Anthropic PBC because of the risks the firm poses to their profits and business models.
DeepSeek’s decision to raise prices could be an inflexion point in China’s AI market, where other top players have followed the company’s playbook in offering low-cost and open-source services. Startups and their investors have been looking for a path to shift the era of ultra-cheap AI services toward more commercially viable business models.
DeepSeek is in the middle of a massive fundraising and has begun preparations for an initial public offering as soon as this year, Bloomberg has reported. Founder Liang Wenfeng will for the first time need to balance investor expectations, rapid market expansion and the intense capital demands of building out costly computing infrastructure.
The company intends to build a massive data centre in Inner Mongolia, part of a plan to secure 1 gigawatt of AI computing power that may include leasing additional capacity from other companies, Bloomberg News reported last week. Nvidia chief executive officer Jensen Huang has previously estimated that a 1GW data centre equipped with the most cutting-edge AI accelerators would require about US$50bil (RM204.47bil), although costs in China are typically lower than in the US.
Customers have many ways to access V4 Flash, including from DeepSeek directly and from other service providers who host the model. Any price increases would hit developers and businesses that pay DeepSeek directly to run its models, and may give rival providers room to lift their own prices. Because V4 Flash’s model weights are openly available under a permissive license, other companies can run it on their own servers and charge customers for access, giving users alternatives if DeepSeek raises its prices too far. – Bloomberg
