The article explores the relationship between AI efficiency and demand, framing it through Jevons Paradox, which suggests that as a resource becomes cheaper, consumption rises. Recent data shows a dramatic surge in AI token consumption, with Google reporting monthly usage hitting 1.3 quadrillion tokens, up from 480 trillion just a year earlier. Despite a slight slowdown in growth—doubling from Q1 to Q2, then increasing by only 33% into Q3—the overall trend remains upward. OpenRouter’s data indicates a similar pattern, with weekly token consumption skyrocketing to just under 6 trillion. Price trends corroborate this demand increase. Token prices have dropped to about a third of their February levels, while consumption has quintupled. This suggests that lower prices are driving higher demand, although the precise relationship remains somewhat unclear. Cloud providers are also seeing incremental revenue growth from AI, with estimates showing that contributions from AI/ML usage are rising from around 1% to 5% quarterly. Fast-growing open-source projects further highlight the integration of AI into development, with six out of the ten fastest projects focused on AI tools. The article raises concerns about the sustainability of this growth, drawing parallels to the shale boom, where initial cash flows dwindled over time. Investors in AI technology are watching closely to see if demand remains robust. Unlike past speculative bubbles, the hope is that efficiencies gained in AI will lead to sustained demand growth, making it a transformative rather than fleeting development in tech.