---
title: "Racing to Sustain Jevons' Paradox"
description: "AI capacity constraints haven't hit spot prices yet, but memory has already doubled. Segmentation into premium, mid-market, \u0026 value tiers is how sellers race to keep Jevons' paradox alive."
categories: ["AI","data"]
keywords: ["GPU pricing","B200 spot market","AI compute costs","model market segmentation","premium mid-market value frontier","HBM3e memory shortage","CoWoS packaging"]
ai_summary: "Hyperscalers declared AI capacity-constrained on Q2 2026 calls. HBM3e memory rose 20% \u0026 HBM4 is forecast to double, while B200 spot rentals have held flat. Model makers are segmenting into premium, mid-market, \u0026 value tiers to keep Jevons' paradox alive through the coming shock."
date: 2026-08-04
lastmod: 2026-08-04
canonical_url: https://www.tomtunguz.com/what-if-gpu-prices-double/
author: "Tomasz Tunguz"
---


There's not enough AI. Jevons' Paradox has been a hallmark of this era like Moore's Law in chips. Will it fail early?

> We continue to be supply constrained, a sign of momentum & rapid adoption. — Sundar Pichai, Alphabet Q2 2026.[^9]

> We will still not have enough capacity to meet all the demand we have in 2026, & I believe this dynamic will also be true in 2027 too. The demand we already have for 2028 is striking. — Andy Jassy, Amazon Q2 2026.[^10]

> Blackwell sales are off the charts, & cloud GPUs are sold out. — Jensen Huang, NVIDIA Q1 FY2027.[^11]

So what happens if the price of AI doubles?

They're already increasing. Anthropic launched Fable 5 on July 24 at $50 per million output tokens, doubling Opus 5 & setting a new frontier ceiling. Google's Gemini flagship climbed from $1.50 to $12 across four generations.[^2]

{{< email_image src="fxldcxggq88bbanwtwdk" alt="OpenAI, Anthropic, & Google flagship output token prices, March 2023 – July 2026, with the July 30 Sol-Luna divergence" width="540" height="391" >}}

Or at least most are increasing. OpenAI cut GPT-5.6 Luna prices by 80% five days after Fable 5's launch, either a market-capture push or a fundamental cost breakthrough.[^12]

The pitch so far from AI : Jevons' paradox drives more consumption as prices fall. But prices are increasing.

AI labs are betting on segmentation. As long as the value & mid-market frontiers absorb the workloads priced out of premium, total GPU-hours consumed keeps growing. Segmentation sustains Jevons.

{{< email_image src="eefqy7ndklswflmtfbzf" alt="Frontier models cost 13× more for 20% more intelligence: premium, mid-market, & value tiers" width="540" height="383" >}}

The premium tier costs 13x the value tier for a fifth more intelligence. The mid-market frontier is the new arrival : GPT-5.6 Sol & Kimi K3, delivering 96% of frontier intelligence at 40% of the cost.[^7][^8] The value frontier runs from GLM-5.2 down to DeepSeek V4 Flash, offering 84% of the intelligence at 1 to 5% of the premium cost.

As segmentation becomes the norm, routers become the strategic layer between buyer & model, the plumbing that lets an application shift a query to the right model. Routers will be internal to models, external in customers' software, & in harnesses.

The major labs each need entries in all three tiers, or they lose the router auction on price. Startups win by owning a single point on the frontier that a big lab cannot economically match. DeepSeek V4 Flash at $0.03 is the existence proof.

All in furtherance of Jevons.

[^1]: [Ornn B200 spot index](https://data.ornn.com/preview) accessed 2026-08-04.

[^2]: [AI Model Inflation](https://tomtunguz.com/ai-model-inflation/). Vendors are pivoting from share-taking subsidies to margin-taking pricing as capex hits records.

[^3]: [Veblen & Jevon Walk Into a Data Center](https://tomtunguz.com/jevons-to-veblen/), March 30, 2026. Token prices dropped 10-20x over 18 months; next-generation frontier models may price up, not down.

[^4]: [Microsoft brought 88 data centers online in FY2026](https://www.datacenterdynamics.com/en/news/microsoft-brought-88-data-centers-online-in-fy2026/), Data Center Dynamics, July 30, 2026.

[^5]: [HBM3e Contract Prices Rise 20% as AI Memory Demand Outstrips Supply](https://siliconanalysts.com/market/hbm3e-contract-prices-rise-20-as-ai-memory-demand-outstrips-supply-2026-06-27), Silicon Analysts, June 27, 2026.

[^6]: [HBM4 Prices to Double Next Year as Samsung, SK hynix Keep Upper Hand](https://en.sedaily.com/finance/2026/07/12/hbm4-prices-to-double-next-year-as-samsung-sk-hynix-keep), Seoul Economic Daily, July 12, 2026.

[^7]: [Artificial Analysis Intelligence Index v4.1](https://artificialanalysis.ai/) accessed 2026-08-04.

[^8]: Mid-market average of GPT-5.6 Sol & Kimi K3 intelligence (58) versus premium average of Opus 5 & Fable 5 (60.5) = 96%. Cost ratio $1.05 / $2.75 = 38%.

[^9]: [Alphabet Q2 2026 earnings call, Sundar Pichai remarks](https://blog.google/company-news/inside-google/message-ceo/alphabet-earnings-q2-2026/), July 22, 2026.

[^10]: [Andy Jassy said Amazon will spend $220 billion this year and still won't have enough capacity to meet demand](https://fortune.com/2026/07/30/andy-jassy-amazon-capex-demand-aws-pga-tour/), Fortune, July 30, 2026.

[^11]: [NVIDIA Q1 FY2027 earnings: revenue, EPS & outlook](https://intellectia.ai/blog/nvda-stock-earnings-analysis-may-2026), May 20, 2026.

[^12]: GPT-5.6 pricing effective July 30, 2026 : Sol at $5/$30, Luna at $0.20/$1.20 per million tokens. [OpenAI GPT-5.6 price-performance frontier announcement](https://openai.com/index/advancing-the-price-performance-frontier-with-gpt-5-6/); [VentureBeat : OpenAI cuts GPT-5.6 Luna prices by 80%](https://venturebeat.com/technology/ai-price-wars-openai-cuts-gpt-5-6-luna-prices-by-80-as-model-competition-shifts-toward-cost).