Soaring Bills Push Businesses Shifts to Cheaper AI Models

Soaring costs are forcing global businesses to rethink their reliance on premium, expensive artificial intelligence models. Tech leaders from companies like Microsoft, Palo Alto Networks and Coinbase now argue that smaller with cheaper models are fully capable of handling everyday corporate needs.

Initially the companies encouraged heavy AI use to boost productivity. However, as AI providers shift from flat subscriptions to usage-based pricing and unpredictable bills are severely impacting budgets. For example, Uber reportedly exhausted its 2026 AI budget in just four months due to a rush on AI coding tools forcing management to heavily cap usage.

With Gartner estimating that AI coding costs could surpass an average developer's salary by 2028, businesses are turning to routing tools like OpenRouter. These platforms automatically assign simple tasks to cost-effective systems while reserving premium AI models for complex work.

Consequently, open-source and cheaper alternatives are surging in popularity. Chinese models like DeepSeek now dominate platforms like OpenRouter, offering capabilities close to top U.S. models for just 18 cents per million tokens a fraction of the $4 average charged by premium competitors. While security concerns may slow their adoption in sensitive sectors, experts note these cheaper alternatives are proving to be "90% as good at 10% of the price."

This cost-conscious shift threatens the revenue growth of industry giants like OpenAI and Anthropic, potentially sparking a fierce price war as they prepare for highly anticipated initial public offerings.