Newsletters
Artificial intelligence is becoming increasingly expensive, prompting companies to reassess how they use the technology as the industry moves beyond its rapid growth phase.
Following the launch of ChatGPT, AI companies adopted a familiar Silicon Valley strategy, offering services at low prices to attract users and gain market share.
Kevin Simback of startup incubator Delphi Labs described the period as one of “subsidized intelligence,” with investors effectively covering costs so companies could provide AI services at discounted rates.
“But the tides are beginning to turn,” Simback said, arguing that major AI firms are now under growing pressure to generate profits as industry leaders such as OpenAI and Anthropic seek to go public and attract retail investors later this year.
Industry analysts say one of the main factors behind rising costs is the growing use of AI agents.
Unlike traditional chatbots that primarily answer questions, AI agents can perform tasks such as scheduling appointments, writing software code and managing files.
These systems require significantly more computing power because a single task may involve multiple agents operating simultaneously.
AI companies typically charge customers based on tokens, the units used to measure and process information. Tasks involving AI agents can consume many times more tokens than standard chatbot interactions.
At the same time, demand for the computer chips and data centers needed to support AI applications continues to outpace supply, increasing costs across the industry.
“Especially in developer circles, the cost to use AI for things like coding has grown exponentially,” said Mark Barton of technology consultancy Omniux.
“All the costs are really starting to skyrocket,” he added.
Some companies have become so reliant on AI that analysts have coined the term “tokenmaxxing” to describe excessive use of token-based services.
“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,” said analyst Jack Gold of J.Gold Associates.
Even Meta, which earlier this year encouraged employees to use as many AI tokens as possible as a measure of productivity, has begun to reconsider its approach.
“Nobody should be using AI tools just for the sake of using them,” Chief Technology Officer Andrew Bosworth wrote in a memo to employees, according to The Wall Street Journal.
Uber’s chief operating officer also questioned whether heavy spending on AI had delivered measurable productivity gains.
To reduce expenses, some companies are turning to free, open-source AI models that can be downloaded and operated independently.
While generally less powerful than leading commercial products such as ChatGPT or Anthropic’s Claude, they are often sufficient for specific business needs.
Others are adopting smaller, specialized AI models tailored to industries such as real estate or finance instead of relying on large general-purpose systems.
Companies are also increasingly dividing complex AI tasks into smaller components and assigning each task to the least expensive model capable of completing it.
“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,” said Adrian Balfour of consultancy Enverso.
Analysts say these trends suggest AI may increasingly resemble a commodity market, where cost efficiency and suitability become more important than using the most advanced model available.
However, experts caution that demand for cutting-edge AI systems is unlikely to disappear.
“The most advanced users” will continue to pay for the highest-performing models, said John Belton, a portfolio manager at Gabelli Funds.
“It’s a growing pie,” he added.
Set as preferred source
Set as preferred source
1 min read
1 min read
