The case for the AI boom hinges on demand, yet the primary risk remains that users simply cannot afford the technology. What has changed is the pricing structure. As OpenAI, Anthropic, and xAI roll out new models with significantly reduced costs, usage is increasing dramatically. This trend suggests the market is generating the volume necessary to justify the trillions of dollars in capital spending currently underwritten by investors.
The Price War Accelerates
The competition has shifted from raw intelligence to cost efficiency. OpenAI is releasing GPT-6 Sol and GPT-6 Luna on Tuesday, which the company says cuts costs for top business customers by 50% compared to previous iterations. Anthropic is launching Opus 5.5 the same day, claiming it costs around 40% less to run than Opus 5 while maintaining top-tier intelligence. Elon Musk's xAI released Grok 4.7 on Monday, pitching the model on price-performance. The speed of this undercutting is notable. It took only 24 hours for new options to erode much of xAI's initial price advantage.
The read-through from this rapid price deflation is that the barrier to entry for developers is collapsing. Citations of Jevons Paradox are appearing in financial analysis, a theory which holds that when technology lowers the cost of a resource, consumption increases rather than decreases. Morgan Stanley views competition from Chinese providers as a bullish signal because it captures more customers who can now access AI at lower price points, thereby increasing overall demand for computing power.
The Counterargument
The counterargument centers on margin compression. OpenAI and Anthropic need to retain customers and sustain revenue growth to justify their sky-high valuations. If the price war continues to intensify, the labs may struggle to maintain profitability even as usage spikes. The risk is that the revenue per user drops faster than the total user base grows, leaving the companies with high burn rates and limited paths to profitability. This is not a hypothetical concern; it is the central tension in the current market structure.
Evidence of Rising Demand
Despite the margin pressure, emerging data points to a rising tide for all providers. Citadel Securities found in a report to clients that falling per-token costs are fueling additional usage, leading to increased overall AI spending. This points to higher eventual profits for the labs and the companies providing the computing infrastructure. The evidence is most visible in the open-weight sector. DeepSeek released V4.1 Flash this month, which it says beats its previous flagship on benchmarks while charging less. That model is now number one on OpenRouter's leaderboard with a 172% spike in usage this week. On balance, the data suggests that cheaper models are driving higher total consumption, a dynamic that supports the broader AI investment thesis.