Trending 
Artificial intelligence (AI) is becoming more expensive, forcing companies to rethink how aggressively they adopt the technology that has dominated the tech industry since ChatGPT’s arrival.
For years, AI firms followed a familiar Silicon Valley strategy: offer powerful tools at low prices to attract users quickly. According to Kevin Simback of the startup incubator Delphi Labs, the industry operated in an era of “subsidised intelligence,” in which investors absorbed much of the cost, allowing companies to provide AI services cheaply.
“But the tides are beginning to turn,” Simback warned, as major AI firms now face pressure to generate profits. Companies such as OpenAI and Anthropic are reportedly preparing for a future in which stronger revenue growth matters more, particularly as they pursue broader investor backing.
One major factor behind rising costs is the growing use of AI agents. Unlike standard chatbots that answer questions, AI agents carry out tasks such as coding, scheduling appointments, or managing files. These systems consume significantly more computing resources because a single request can activate multiple agents working simultaneously.
That surge in activity drives up token consumption, the unit AI companies use to bill customers. A complex AI task may consume dozens of times more tokens than a standard chatbot interaction.
The pressure does not stop there. Demand for computer chips and data centres continues to outpace supply, creating computing shortages across the industry.
“Especially in developer circles, the cost to use AI for things like coding has grown exponentially,” said Mark Barton of tech consultancy Omniux. “All the costs are really starting to skyrocket.”
Some organisations have also embraced excessive use of AI in a trend known as “tokenmaxxing.” Analyst Jack Gold of J.Gold Associates said some firms have discovered that AI spending can quickly exceed labour costs. “In some cases people are seeing the cost of tokens exceed the cost of the employee within a month or two of use, just because they’re using it too much,” he said.
Even Meta has reconsidered its approach. After previously encouraging heavy AI use internally, chief technology officer Andrew Bosworth reportedly told staff that “nobody should be using AI tools just for the sake of using them.”
Concerns about returns on AI spending are also spreading. This week, Uber’s chief operating officer questioned whether heavy AI investment was translating into measurable productivity gains.
In response, companies are adopting cost-cutting strategies. Some are turning to free open-source models, while others prefer smaller, specialised systems tailored to industries such as finance or real estate.
Adrian Balfour of consultancy Envorso said costs can vary sharply depending on the model chosen. “The big large monolithic model, it’s $15 per million tokens, but you can get that down to like five cents if you use the smaller mini model,” he said.
The shift suggests AI may increasingly resemble a commodity market, where value depends less on using the most advanced model and more on finding the right tool at the right price.
Still, analysts say premium AI systems are unlikely to disappear. John Belton, a portfolio manager at Gabelli Funds, believes top-tier models will continue attracting high-end users willing to pay for peak performance.
