Why AI Pricing Is Becoming a Major Challenge
Artificial intelligence (AI) has become part of everyday life, with millions of people using free versions of ChatGPT, Gemini, and Claude for writing, coding, and planning tasks.
While these services appear free to users, companies such as Microsoft, Google, and Anthropic have invested billions of dollars to develop the large language models (LLMs) that power them. To recover these costs, firms also offer premium AI subscriptions with advanced features.
However, pricing AI services is proving to be one of the industry’s biggest challenges.
The main reason lies in the use of tokens, the basic units that AI models process when users submit prompts and receive responses. Every question and every AI-generated answer is broken into tokens. The more tokens used, the higher the cost.
Unlike traditional software, AI does not always produce the same result for the same prompt. Small changes in wording can generate different responses and consume different numbers of tokens. The challenge becomes even greater in agentic AI systems, where multiple AI agents work together to complete complex tasks, significantly increasing token usage.
Although the price of individual tokens has dropped in recent years, overall token consumption is rising rapidly. Goldman Sachs estimates that global token usage could increase 24 times between 2026 and 2030 as businesses increasingly adopt AI agents.
Many companies are struggling to predict how many tokens their employees or customers will use. Reports suggest that some organizations have exhausted their AI budgets much earlier than expected because staff relied heavily on AI-powered coding and productivity tools.
Experts say this unpredictability makes financial planning difficult. Simon Gooch of identity management company Saviynt believes it is almost impossible to create long-term pricing models because AI costs continue to change. Similarly, Professor Will Venters of the London School of Economics notes that businesses often find it hard to control spending because AI outputs—and therefore costs—are not fixed.
Some companies are trying to reduce expenses by using cheaper AI models or writing more detailed prompts to limit unnecessary token usage. Others have relied on low-cost personal AI subscriptions, although industry experts believe major AI providers may eventually restrict such practices as they seek stronger profits.
Pricing becomes even more complicated when businesses integrate AI into products used by thousands of customers. Token consumption can grow rapidly as companies use AI not only for software development but also for testing, security, customer support, and automated decision-making.
Despite the rising costs, experts argue that AI can still deliver strong value by improving productivity and automating complex tasks. The challenge is finding a fair way to charge customers for these benefits.
Software companies are currently exploring several pricing options, including higher subscription fees, charging based on results, or offering bundles of AI-powered services. However, these models remain uncertain because AI providers frequently adjust their own pricing structures.
As AI adoption continues to expand, businesses face a difficult balancing act: controlling unpredictable token costs while offering affordable and profitable AI services. For now, industry experts agree that there is no perfect pricing model, and companies are still searching for a sustainable solution.
About The Author
