Jensen Huang’s AI paradox: Cheaper tokens fuel demand, but GPU costs stay high
NVIDIA CEO Jensen Huang has framed falling AI token prices as a driver of demand, likening the trend to Jevons’ paradox. However, data shows H100 GPU rental prices remain stable or rise, even as token costs drop. The dynamic raises questions about the sustainability of the AI ecosystem’s growth assumptions.
Editor, Lazyfounder
30 SEC SUMMARY
- NVIDIA CEO Jensen Huang argues that falling AI token prices are driving increased demand, mirroring Jevons’ paradox.
- Data from Ornn, Silicon Data, and Bloomberg shows H100 GPU rental prices remain stable or rise despite cheaper tokens.
- Cheaper AI tokens enable new applications like AI agents, which consume tokens at a high rate and may inflate compute demand.
- The AI ecosystem’s stability depends on continued growth in usage to offset declining token prices.
- US stocks dipped after reports suggested OpenAI’s annualized revenue fell short of expectations.
TABLE OF CONTENTS
KEY HIGHLIGHTS
- NVIDIA CEO Jensen Huang says cheaper AI tokens are fueling demand, akin to Jevons’ paradox in the AI market.
- H100 GPU rental prices have held steady or increased despite falling token prices, per data from Ornn, Silicon Data, and Bloomberg (August 2026).
- Agentic AI systems consume tokens rapidly, potentially inflating compute demand and creating new use cases.
- The AI ecosystem relies on sustained growth in AI usage to balance declining token prices and maintain stability.
- US stocks reacted negatively to reports of lower-than-expected annualized revenue for OpenAI.
Cheaper AI tokens drive demand, says NVIDIA CEO
Jensen Huang, CEO of NVIDIA, has framed falling AI token prices as a catalyst for increased demand, likening the trend to Jevons’ paradox, according to The Decoder. In economics, Jevons’ paradox describes a scenario where improved efficiency or lower costs of a resource lead to greater consumption rather than conservation. Huang’s argument suggests that as AI tokens become cheaper, demand for AI applications—and consequently, compute resources—will rise.
This perspective aligns with data from Ornn, Silicon Data, and Bloomberg, which shows that while token prices have declined, rental prices for NVIDIA’s H100 GPUs have remained stable or even increased as of August 2026. The divergence highlights a growing tension in the AI market: lower token costs are enabling new use cases, but hardware scarcity and sustained demand are keeping compute costs high.
What this means
Lazyfounder analysis — our interpretation, not reported fact.
Jensen Huang’s perspective on falling AI token prices is a double-edged sword for founders and operators. On one hand, cheaper tokens lower the barrier to entry for AI applications, enabling startups to experiment with agentic systems, automation, and other high-token-use cases. This could accelerate innovation and adoption, which is a net positive for the ecosystem.
On the other hand, the dependence on sustained demand growth to offset falling token prices introduces a fragile dynamic. If adoption plateaus or AI applications fail to scale as quickly as token prices decline, the entire supply chain—from hardware manufacturers to cloud providers—could face revenue pressures. This risks creating a boom-and-bust cycle, where overinvestment in compute infrastructure meets softened demand.
For founders, the takeaway is clear: build for efficiency, not just scale. Startups that optimize token usage, reduce waste, and deliver tangible value per token consumed will be better positioned to weather volatility. Operators should also prepare for scenarios where compute costs remain high even as token prices fall, as GPU rental prices show no signs of dropping. The AI market’s growth may be nonlinear, and those who plan for both expansion and contraction will fare best.
Key takeaways
- Founders building AI-driven products should anticipate persistent demand for compute resources, even as token prices fall.
- Startups relying on AI agents or automation must account for high token consumption and its impact on operational costs.
- The stability of the AI supply chain—from chipmakers to energy providers—depends on continuous growth in AI adoption.
- Investors and operators should monitor revenue growth closely, as market reactions to missed targets can ripple across the sector.
FAQ
What is Jevons’ paradox, and how does it apply to AI?
Jevons’ paradox is an economic phenomenon where increased efficiency or lower costs of a resource lead to greater consumption rather than reduced usage. In the AI market, this means that as token prices fall, demand for AI applications—and the compute resources they require—could rise, rather than staying flat or declining.
Why are H100 GPU rental prices not dropping despite cheaper tokens?
H100 GPU rental prices remain stable or rise because demand for compute resources is outpacing supply. Even as token prices fall, the emergence of new AI applications like agentic systems and automation is driving sustained or increased demand for GPUs, keeping rental costs high.
What happens if AI demand doesn’t grow fast enough to offset falling token prices?
If AI usage fails to grow at a rate sufficient to counterbalance falling token prices, the entire AI supply chain—from chipmakers to energy providers—could face financial strain. Revenue streams may shrink, leading to reduced investments in infrastructure and capacity, which could slow down innovation and adoption.
Related on Lazyfounder
Sources
- The Decoder · 2026-10-11
Cheaper AI tokens are driving more demand, and that's Jensen Huang's best-case scenario
This story is an original summary drafted with AI by Lazyfounder from the reporting listed above and checked by automated validation. Facts are attributed to their original publishers; sections marked as analysis are Lazyfounder's. Where a source is in another language, facts were machine-translated and quotations are reported, not reproduced. Read the original coverage via the links, and see our AI policy and corrections policy.
About the author
Editor, Lazyfounder
Tarun Mottlia edits LazyFounders, covering Indian startups, funding rounds, AI and product launches. Every story on the site is AI-assisted and checked against its cited sources before publication.
More stories by Tarun MottliaGet the LazyFounder Brief
Startup, funding and AI news in a five-minute read. Join the early-access list.
