Why charging for AI usage remains a challenge

Companies invest billions in large language models, but still struggle to turn that cost into stable revenue.

On August 3, 2026, experts highlight that the free use of assistants like ChatGPT still provides a huge benefit, even as the costs behind the technology soar.

Massive investment in large language models

Companies such as Microsoft, Google, and Anthropic have poured hundreds of billions of dollars into developing large language models (LLMs), the technology that powers both free and paid AI services.

How tokens power AI

When a user asks an LLM, such as ChatGPT or Claude, for a response, the prompt is broken down into mathematical blocks called tokens. The model’s answer also arrives as tokens, which are then converted back into text, code, or commands.

Explosion in token consumption

According to a Goldman Sachs analysis, although the per-token price has fallen in recent years, the volume consumed by businesses and consumers has skyrocketed. The bank projects that external token consumption will increase 24-fold between 2026 and 2030, reaching 120 quadrillion tokens per month as organizations adopt AI agents.

Spending management challenges

Companies and individuals often lose track of how many tokens they are using until they receive a bill or run out of credits. "Trying to tie someone into a cost model for the next 12 months, two years, three years, it doesn't make any sense, honestly, because we don't know," says Simon Gooch of Saviynt. Will Venters of the London School of Economics warns that teams can burn tokens while experimenting with AI internally: "People are finding it really hard to manage that cost… it's a non-deterministic output, so it's a non-deterministic value." Rob Steele, CFO of iplicit, likens the need for detailed instructions to sending someone to the market without a shopping list.

Pricing models and shareholder pressure

Oliver King-Smith, founder of smartR AI, notes that smaller organizations may use personal fixed-rate accounts, something big providers don’t like, and predicts that "This has to end at some point in time, because the big guys are taking a bath on those accounts." When shareholders demand profit, the major platforms will likely tighten the reins. Bill Peterson of Sumo Logic remarks that the choice of pricing structure can shift if LLM providers change their own rates: "You get into variable pricing, and it's changing every couple of months. Customers don't like that. That's not how anybody builds a budget."

With information from BBC News.