The Silicon Data LLM Token Expenditure Index has declined nearly 20% from its May high after nearly doubling since its inception in December, according to reports from The Hindu BusinessLine. This decline is raising concerns among investors about whether the enormous capital investments being poured into artificial intelligence will ever generate sufficient returns. The index tracks what users pay for AI tokens and serves as a key indicator of the $700 billion-plus capex boom that has been driving the sector's growth. However, recent analysis from Exponential View reveals that token demand appears elastic: as prices fall, usage grows faster, with the magnitude of elasticity approximately 1.2-1.8: every 10% price cut → 12–18% more tokens. As reported by Exponential View, the cost to use a given level of AI falls about 10x every 12 months, and lower prices lead to much more use, with Google processing 9.7 trillion tokens a month, now over 480 trillion — 50x more.
Veteran investor Louis Navellier reports increasing concerns about AI companies losing pricing power as users of token-priced AI solutions face cost constraints. As reported by The Hindu BusinessLine, there are growing reports that users are having to restrain unlimited usage due to high costs. The chatter that OpenAI is pushing back its IPO to next year is being interpreted as a sign that current profitability remains problematic for the sector. According to Exponential View, the cost to use a given level of AI falls about 10x every 12 months, and lower prices lead to much more use, with Google processing 9.7 trillion tokens a month, now over 480 trillion — 50x more. Recent data shows that Doubao's daily token usage exceeded 50 trillion this month, up from 4 trillion in Dec 2024, demonstrating the dramatic scale of usage growth as prices decline. Investing.com analysis suggests the problem may be that customers may be shifting toward cheaper models or discovering that unlimited AI usage looks a lot less attractive once the monthly bill arrives.
Recent regulatory developments are creating additional pressure on AI pricing. According to The Hindu BusinessLine, the US government removed foreign access restrictions on Anthropic PBC's Fable 5 model this week, while regulators requested OpenAI to stagger the roll-out of an upcoming release. The European Union's AI Act targets frontier models for mandatory evaluations and stringent transparency requirements, creating deployment-and-compliance burdens that may push companies toward cheaper models. Exponential View analysis shows that the transition from chat to agents is multiplying token use, with agent coordination density increasing and more capable models able to cover a wider range of economically useful tasks. This regulatory environment is creating a complex landscape where pricing power is being challenged by both market dynamics and regulatory requirements.
Despite the pricing concerns, there are mixed signals in the market. As reported by The Hindu BusinessLine, while token prices have collapsed more than 90% since 2023, total spend has roughly doubled since last year, indicating that cheaper tokens are expanding the market. However, Allianz Research warns of a 46% growth gap between AI investment and sales, which is worse than the 32% divergence measured during the 2001 telecom bust. This creates uncertainty about whether the current capex spending justifies returns, particularly as top-end graphics processing units and high bandwidth memory remain sold out through 2026. Investing.com analysis suggests that the likely next phase is not necessarily a chip glut, but a rotation from premium training demand toward cheaper, more practical inference workloads. The analysis reveals that global token volumes exceed 30 quadrillion per month, growing 14x year-over-year, demonstrating the unprecedented scale of AI adoption despite pricing pressures.
