“I won’t be surprised if over the course ​of this year, companies will be really scrutinising [the cost of AI],” Comyn said.

Data centre costs and workforce disruption have pushed the price of AI tokens, which previously were quite cheap, up, alongside the fact that these AI tools are getting more advanced and have greater “reasoning, the access to tools, the amount of context that you can put into it – your token costs do not scale on a linear basis,” Comyn added.

AI giants are also shifting from flat subscriptions to usage-based pricing, meaning companies are ending up with higher bills despite AI promising cheaper operational costs than human workers.

“Changing the licence model caught a lot of people by surprise,” said Harold Byun, CEO of AI implementation start-up, BlueRock.

“Immediately after that, we had a number of reports from customers that we’re seeing a 20 to 30 per cent spike in terms of over-budgeting.”

Rideshare and technology company Uber, for example, was forced to limit AI usage after it depleted its entire 2026 AI budget in just 4 months, after staff quickly adopted AI coding tools.

Palo Alto CEO Arora has urged AI firms to cut pricing, and allow businesses to buy tokens in bulk at the lower rates to use in the future.

“If you want to win enterprise, you should be forward pricing tokens. The cheaper the tokens for enterprises it will allow for experimentation, workflow reimagination - instead CIOs are busy restricting AI use and working on making the use more efficient!

Meanwhile, according to Gartner, AI coding costs are estimated to surpass that of an average developer’s salary by 2028. As a result, as tech giants are arguing, cheaper models such as OpenRouter are increasing in popularity, with the number of tokens processed increasing from 34 per cent in January to 65 per cent in June, according to Citi.

This could also increase appeal in Chinese models, which historically had capabilities far behind their American rivals, but are quickly catching up.

OpenRouter’s top 4 models are all Chinese, including DeepSeek at number 1.

Chief AI officer at Weka, Val Bercovici, said that these open-source models are “90 per cent as good at 10 per cent of the price.

“We don’t need to spend the premium tokens on every level of effort.”

The other impact the appeal of these cheaper models has is on the profit margin (or lack thereof) of US AI giants like OpenAI and Anthropic, both of which are looking to secure Initial Public Offerings (IPOs) and be first to the public market.

OpenAI is considering price cuts, including on tokens, just as Anthropic is believed to be.

However, these lower costs could impact revenue growth as IPOs approach.

“There will be a price-war dynamic when it comes to OpenAI and Anthropic as they both duke it out for a ‘first to public market’ IPO dates,” said financial adviser for private wealth management at Synovus Securities.