Editorial illustration for Cheaper AI Tokens Drive More Demand, a Best Case for Jensen Huang
Cheap AI Tokens Boost Demand, Nvidia Profits Soar
Cheaper AI Tokens Drive More Demand, a Best Case for Jensen Huang
Token prices for AI models have been sliding for two years straight, yet Nvidia's H100 rental rates refuse to follow. That gap is the whole story right now. Data compiled by Ornn, Silicon Data, and Bloomberg through August 2026 shows a pattern the venture firm a16z describes as a textbook case of Jevons' paradox: when something gets cheaper, people don't use less of it, they use so much more that total demand goes up, not down.
For Jensen Huang and Nvidia, this is the scenario they've been betting on since the ChatGPT boom started. Cheap tokens make AI agents and automated workflows economically viable in ways they weren't a year ago, which pulls in more usage, which keeps GPU capacity tight even as the per-token cost collapses. The catch is nobody can cleanly separate human-driven demand from machines querying machines, and agentic systems chew through tokens at rates that dwarf anything a person typing into a chatbot produces. That ambiguity matters, because it determines whether the hardware shortage propping up chipmakers, memory suppliers, and cloud providers reflects real growth or something more fragile.
Data from Ornn, Silicon Data, and Bloomberg (as of August 2026) shows what a16z calls a textbook Jevons paradox in the AI market. Token prices keep dropping, but H100 GPU rental prices hold steady or climb. Cheaper tokens unlock AI agents, automation, and new applications, so volume grows faster than per-unit costs fall.
Why this matters
For Nvidia, this is the dream outcome: cheaper tokens didn't kill the GPU business, they expanded it. H100 rental prices holding steady while token costs fall is the clearest signal yet that demand, not just capacity, is driving the market. But we'd push back on the easy read here.
The data can't separate human-driven demand from agentic systems calling APIs in loops, and that distinction matters a lot for anyone pricing infrastructure or planning product roadmaps. If most of the new volume is agents burning tokens to accomplish tasks rather than people doing new things, the Jevons framing starts to look less like organic market growth and more like a measurement artifact. Founders building on top of these token economics should ask whose usage is actually paying the bills.
Researchers tracking compute trends should want the human-versus-agent split before citing this as proof that cheaper AI reliably creates proportionally larger markets. Right now, that's assumed, not shown.
Related Reading
AI Tools & Apps Claude gains shared context in Excel, PowerPoint; Microsoft adds Copilot Cowork
AI Tools & Apps OpenAI Developer: AI Agent "Swarms" Waste Tokens, Create "Coordination Tax
AI Tools & Apps LG's recent webOS update adds Microsoft Copilot app, now removable
From Archives Meta's AI Logo Designer Faces Backlash After Revealing Workshop Fee
From Archives Microsoft's Copilot Gains File Access in Hybrid Intelligence Demo
Common Questions Answered
What is Jevons' paradox and how does it apply to AI token pricing?
Jevons' paradox describes a phenomenon where cheaper prices for a product lead to increased overall demand rather than decreased usage. In the AI market, as token prices have dropped over two years, users have dramatically increased their consumption of AI services, causing total demand to rise instead of fall, which benefits GPU providers like Nvidia.
Why are H100 GPU rental rates staying high despite falling AI token prices?
H100 rental rates have remained steady or climbed because cheaper tokens have unlocked new AI applications, automation systems, and AI agents that require massive computational resources. The increased volume of AI workloads driven by lower per-token costs means demand for GPUs continues to grow, preventing rental prices from declining alongside token costs.
How does the gap between token price declines and H100 rental rates benefit Nvidia and Jensen Huang?
This gap represents Nvidia's ideal scenario because it demonstrates that cheaper tokens are expanding the GPU market rather than shrinking it. Instead of lower token prices destroying demand for expensive hardware, they're driving more consumption of computational resources, allowing Nvidia to maintain high rental rates while the overall AI infrastructure market grows.
What data sources were used to identify the Jevons' paradox pattern in AI markets?
According to the article, data compiled by Ornn, Silicon Data, and Bloomberg through August 2026 was used to document the pattern of falling token prices paired with steady or rising H100 rental rates. This data was analyzed by venture firm a16z, which identified it as a textbook case of Jevons' paradox in the AI market.
What uncertainty exists in interpreting the relationship between token prices and GPU demand?
The article notes that available data cannot clearly distinguish between human-driven demand for AI services and automated agentic systems making repeated API calls. This distinction is critical because it affects how infrastructure providers and product teams should interpret pricing trends and plan their roadmaps for the future.
Further Reading
- Cheaper AI tokens are driving more demand, and that's Jensen Huang's best-case scenario - The Decoder
- 35x Lower Token Cost with Blackwell - NVIDIA
- AI Demand Diverges: Token Usage Soars, H100 Falls While B200 Rebounds, DRAM Drops After Five Straight Gains - BTCC
- AI Inference Cost Statistics (2026): 50+ Data Points on the Cost of Running AI - VoxBooster
