Anthropic's most advanced AI model, Fable 5, is barely being adopted by U.S. companies despite being considered the most capable AI system available. New spending data reveals that the model accounted for just 11.4% of Anthropic's total token spending in its first month after launch, raising questions about whether the AI industry's growth strategy is sustainable.

Why Isn't Fable 5 Taking Off Despite Its Superior Performance?

The numbers tell a striking story. Fable 5 generated only about 75% of the revenue that OpenAI's flagship GPT-5.6 Sol model produced, even though Fable 5 costs roughly twice as much per token to use. According to spending data from financial services provider Ramp, Fable 5 costs about $10 per million input tokens and $50 per million output tokens, making it significantly more expensive than competing models.

The performance gap between Fable 5 and cheaper alternatives may not justify the cost for most business use cases. Many companies struggle to measure the real-world value that incremental improvements in AI capability deliver to their operations. When the difference between one model generation and the next is difficult to quantify in daily work, the willingness to pay sharply higher prices appears to have limits.

"Price may be hitting a ceiling for corporate AI spending," noted Ara Kharazian, an economist at Ramp, explaining that the extra performance simply isn't worth the cost for most organizations.

Ara Kharazian, Economist at Ramp

However, the issue may be more nuanced than price alone. The real challenge is calculating return on investment for AI tools. How does a company assign a dollar value to productivity gains from a language model? How do you measure the gap between one model generation and the next in concrete business terms? These questions remain largely unanswered, leaving companies cautious about spending premium prices on marginal improvements.

What Does Fable 5's Slow Adoption Tell Us About the Broader AI Market?

Fable 5's underwhelming performance has broader implications for the AI industry's investment thesis. The sector has been built on the assumption that companies will consistently pay more for increasingly powerful models. If that assumption breaks down, it threatens the financial models of leading AI labs that depend on growing revenue from premium offerings.

The spending patterns reveal a market in transition. According to Ramp's data, 43.5% of U.S. companies paid for Anthropic subscriptions or tokens in July, up 1.1 percentage points from the previous month. OpenAI reached 39.7% of companies but grew by only 0.23 percentage points, significantly lagging overall AI adoption growth. Meanwhile, xAI posted its fastest growth since July 2025, rising 0.94 percentage points to 4%.

The most concerning trend for industry leaders involves advanced users. Those who spend the most on AI are increasingly shifting toward open-source models, which now trail frontier models by only a few months in capability. This shift is slowing growth at both OpenAI and Anthropic, the two leading AI labs.

How Companies Are Actually Spending on AI

• Top 1% of companies: Spent a median of $7,400 per employee on AI in July, representing the highest tier of AI investment
• Top 10% of companies: Spent $650 per employee on average, showing significant variation in spending levels across high-adoption organizations
• Median company: Spent $11.95 per employee on AI, indicating that most organizations are still in early stages of AI deployment

Despite the skepticism around premium pricing, total AI spending continues to rise across the board. Companies are investing more in AI overall, but they're becoming more selective about which models they use and how much they're willing to pay for incremental improvements. The data suggests a market maturing from early enthusiasm into more disciplined purchasing decisions.

The fundamental question facing the AI industry is whether models that are dramatically more capable can deliver dramatically higher and tangible value. Companies will pay for AI tools that clearly improve their bottom line. But as long as the value proposition remains abstract or difficult to measure, their willingness to pay for premium pricing appears to have plateaued. For Anthropic and other frontier AI labs, converting capability into demonstrable business value may be the next critical challenge.