China’s low-cost AI surge emerges as an IPO wildcard for OpenAI and Anthropic
Low-cost Chinese AI models are rapidly increasing their share of enterprise AI traffic, raising the possibility that the IPO valuation models for OpenAI and Anthropic could be unsettled. Artifact Analysis estimated that Claude was about 9 times more expensive than low-cost Chinese models, while Google said switching models could save more than 10hundreds of millions of dollars a year.
Low-cost Chinese artificial intelligence models are rapidly increasing their share of enterprise AI traffic. Observers say this is weakening the technological advantage that high-priced U.S. models have held in the enterprise market. The shift is accelerating in tandem with the cost burden on corporate customers. In recent earnings announcements, Meta, Spotify, and Pinterest said that rising AI-related costs were weighing on profitability. More companies are directly citing AI costs in their earnings disclosures and revealing the financial burden. Repeated reports confirm that rising costs are affecting companies’ operating performance. The expected IPO valuations of OpenAI and Anthropic have been based on market share and pricing power. Their valuations rest on the assumption that customers will pay a premium because they have no alternatives. Analysts say the intensifying cost competition could challenge that assumption. The number of cutting-edge frontier AI models is also growing, while costs are trending lower. Chinese research labs are said to offer similar capabilities at lower cost. Western companies such as Mistral and Cohere are also rolling out cheaper alternatives one after another. AI benchmarking firm Artifact Analysis estimated that Anthropic’s Claude was about 9 times more expensive than the cheapest Chinese model for the same task. The comparison has been presented as a benchmark for companies comparing costs. The figure is being used as a reference in corporate decision-making. Google said that if a major customer of its cloud service switched 80% of its workload from frontier models to “Gemini 3.5 Flash,” it could save more than 10hundreds of millions of dollars a year. Google cited the potential for cost savings through model switching as an example. The scale of the savings has drawn attention, based on the case of a major customer. These developments have prompted discussion that the components underpinning IPO expectations could change. Analysts point out that a loss of pricing advantage could affect how companies’ valuations are estimated. The issue is being discussed as a factor in potential IPOs.
Sources: 뉴시스 경제
Bullstory synthesized this report from the linked sources. AI may have assisted the writing process. Check the original reports for factual confirmation.