The global artificial intelligence market has entered a fierce pricing contest. Rising corporate technology bills are forcing businesses to closely scrutinize their operational expenditures, prompting customers to curb heavy usage and seek affordable alternatives. This cost-conscious shift has allowed Chinese developers including Moonshot and DeepSeek to make significant inroads with users ranging from Silicon Valley to Europe.
Jefferies Data Shows Enterprise Inference Costs Plummeting to Yearly Lows
The cost for businesses to run artificial intelligence models has dropped to a yearly low, according to research published by investment bank Jefferies. Average inference prices—tracked per million tokens, or chunks of data handled by a model—ranged between US$1.16 and US$1.18 from August 6 to 8. Jefferies noted that this marked the lowest level recorded this year, citing data from US research firm Silicon Data.
Those figures demonstrate a sharp contraction from earlier in the year. Average unit costs have plummeted from US$2.04 on May 31 and US$1.45 in late July. The index compiled by Silicon Data tracks pricing across business application programming interface providers and open-weight inference platforms utilized by software developers.
OpenAI and Anthropic Respond With Aggressive Model Rate Cuts
In response to mounting competitive pressure from lower-cost alternatives, leading US artificial intelligence laboratories are rolling out heavily discounted services. Jefferies analysts led by Thomas Chong pointed out that the price decline coincided with an increasing emphasis on cost efficiencies
across both the United States and Chinese technology ecosystems.
To defend their market share against affordable open-source tools, major American firms have moved quickly to slash their rates. OpenAI escalated the price war by cutting rates for its latest GPT-5.6 model series by up to 80 per cent. At the same time, Anthropic introduced its Claude Opus 5 model, delivering performance comparable to its flagship Fable 5 model at half the price.
Open-Source Pioneers Drive Down Computing Expenses
While proprietary US labs adjust their pricing strategies to retain clients, open-source developers continue to push the boundaries of affordable computing. Platforms originating in China are altering the economic calculus for enterprise software developers by providing high-performance capabilities at a fraction of traditional proprietary costs. DeepSeek, for instance, offers tools that help drive down expenses.
These pricing structures reflect a broader historical compression in the cost of artificial intelligence deployment. Historical figures indicate that the cost of querying models has fallen over time. As capital expenditures by major technology firms continue to scale, the race toward low-cost inference is reshaping how global enterprises procure and deploy software infrastructure.
