The artificial intelligence industry is currently experiencing a significant price war, with leading US firms like OpenAI and Anthropic slashing the costs of their models. This aggressive pricing strategy, driven by the emergence of cheaper Chinese competitors such as DeepSeek, Moonshot, and Zhipu, has led to substantial reductions. OpenAI, for instance, cut the price of its lightweight GPT-5.6 Luna model by 80% to $0.20 per million input tokens and $1.20 per million output tokens, and its mid-tier Terra model by 20%. Anthropic responded by launching Claude Opus 5 at roughly half the cost of its predecessor, Fable 5, aiming to offer frontier-level performance at a more competitive price point.
This price competition is forcing AI companies to prioritize market share and user growth over immediate profitability, a strategy known as 'blitzscaling.' The goal is to establish dominance by offering products at below cost, hoping to monetize later. For example, running a task on Zhipu's GLM model can cost around $544, compared to $4,811 for the same task on Claude, highlighting a nearly ninefold price gap. This dramatic difference is now influencing enterprise purchasing decisions, making it harder for US firms to justify their premium pricing.
Analysts note that the average prices for AI inference from leading US labs fell nearly 25% between mid-July and mid-August. This rapid decline is much steeper than the historical trend of around 57% per year, pushing AI from a software-driven business to a capital-intensive one where competitive advantage increasingly relies on balance sheets rather than just model sophistication. The price war also has implications for data center demand, as AI labs are more likely to seek flexible, scalable capacity commitments rather than long-term fixed leases, and will prioritize power and cooling availability as key constraints. Investors are scrutinizing whether increased volume can offset thinner margins to justify trillion-dollar valuations.
