Business & Finance
AI token index falls to record low amid growing market competition
Declining token costs could reduce operating expenses for AI users but squeeze revenues for providers like OpenAI and Anthropic.
The short version
- The LLM Token Expenditure Index from Silicon Data fell to 97 cents on Monday, its lowest reading since its launch late last year.
- Price declines are being driven by lower production costs, price cuts by frontier labs, and the rise of cheap open-source models like China's Kimi K3.
- The deflationary trend could pressure margins for AI developers like OpenAI and Anthropic, both of which reportedly filed confidentially for IPOs this summer.
Key facts
- The LLM Token Expenditure Index, which tracks the daily market rate for large-language model tokens, dropped to 97 cents on Monday, down more than 50% from its summer high.[CNBC]
- The price drop is attributed to decreasing production costs, dynamic pricing by some labs, and rising competition from open-source Chinese models like Moonshot's Kimi K3.[CNBC]
- OpenAI announced price cuts for two of its GPT-5.6 models in late July.[CNBC]
- Both Anthropic and OpenAI confidentially filed for initial public offerings with regulators this summer.[CNBC]
What remains uncertain
- It remains to be seen how shifting token prices will impact long-term investor returns for megacap tech firms that have invested billions in AI infrastructure.[CNBC]