Model launches

AI-generated text

Falling prices for powerful AI models drive usage growth and bolster AI boom outlook

Major AI developers are cutting operational costs for their top models, leading to sharply higher usage.

Falling prices for powerful AI models drive usage growth and bolster AI boom outlook

For years AI companies competed to build the best — and often riskiest — models. This week attention has shifted toward cutting costs: several firms released new models that deliver high intelligence at prices that would have been hard to imagine months ago.

Why it matters

The main threat to continued AI growth is weak demand. Innovations that lower model costs while preserving capability provide a bullish signal: cheaper access boosts usage, which helps justify the large-scale investments the industry anticipates.

Current situation

This week saw multiple major releases from OpenAI, Anthropic, xAI (Elon Musk) and several Chinese providers. Common to these launches is a significant reduction in operating costs for top-tier models.

  • OpenAI is releasing GPT-6 Sol and GPT-6 Luna on Tuesday; the company says these reduce costs for top business customers by 50% versus prior iterations in the same family.
  • Anthropic is releasing Opus 5.5 on Tuesday; the company reports it costs roughly 40% less to run than Opus 5 while maintaining top-level intelligence.
  • xAI released Grok 4.7 on Monday, positioning the model around price-performance, though rapid competitive moves have quickly eroded some of its initial price advantage.

Risks and competitive pressure

An ongoing and rapid price war raises concerns for OpenAI and Anthropic, which need to retain customers and revenue growth to support their multi‑trillion‑dollar valuations. Despite margin pressure, the trend of falling costs is overall positive for the AI boom because it makes broader adoption easier to justify.

Financial signals

There is emerging evidence that lower unit costs are increasing usage and could raise revenues for companies providing AI compute. Citadel Securities told clients that declining per‑token costs are fueling additional consumption and rising overall AI spend. Morgan Stanley views competition from Chinese providers as another bullish indicator, since lower prices can expand the user base and increase total demand for compute.

Economists refer to this dynamic as the Jevons Paradox: when technology reduces the cost of a resource, consumption can rise.

Open‑weight models and leaderboard shifts

Open‑weight model vendors, mainly from China, have advanced rapidly. DeepSeek’s V4.1 Flash, released this month with technical updates, reportedly outperforms its previous flagship on several benchmarks and allowed the company to charge less. That model climbed to number one on OpenRouter’s leaderboard, showing a 172% spike in usage this week.

A reality check

Using lower prices to capture market share is not new in AI. Firms have historically released new models, discounted older releases, or issued smaller, cheaper “flash” versions. What has changed is the scale and speed of competition from open‑weight providers, which is putting broader pressure on pricing across offerings from leading AI labs.

Bottom line

The sticker shock around AI costs is receding as labs push for cheaper models. Falling prices and rising usage together strengthen the case for continued expansion of AI, even as companies must navigate fierce competition to maintain revenue and valuation.