After the AI binge, companies balk at soaring bills

The initial frenzy of AI adoption is giving way to a more cautious approach as companies grapple with soaring operational costs. Following the widespread excitement around AI, many firms initially offered services at heavily discounted rates, a strategy described as "subsidized intelligence" where investors bore the brunt of the expenses to capture market share.

“

But the tides are beginning to turn, and an era where the big AI companies actually need to make money has begun.

However, this era of cheap AI is ending. Leading companies like OpenAI and Anthropic are preparing for public offerings, signaling a shift towards profitability. Prices for AI services are climbing across the board, significantly impacted by the rise of AI agents. These agents go beyond simple chatbots, performing tasks like booking appointments or writing code, but their complex operations can consume vast amounts of computing resources and tokens, the unit by which AI usage is billed.

“

Especially in developer circles, the cost to use AI for things like coding has grown exponentially. All the costs are really starting to skyrocket.

One major factor driving up costs is the intense demand for the specialized computer chips and data centers required to power AI. This demand has created significant shortages, adding to the uncertainty and expense within the nascent industry. "The cost to use AI for things like coding has grown exponentially," noted Mark Barton of tech consultancy Omniux, adding that "all the costs are really starting to skyrocket."

“

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.

Some companies, having engaged in what's termed "tokenmaxxing" – excessive AI usage – are now seeking to curb spending. Meta, which previously encouraged employees to maximize token usage, has since issued a memo advising against using AI tools merely for the sake of it. Uber's chief operating officer even questioned the productivity gains from current AI spending. In response, businesses are exploring alternatives such as free, open-source AI models, which, while less powerful than premium options, are sufficient for many tasks. Others are opting for smaller, specialized AI models tailored to specific industries, or breaking down large AI tasks into smaller, more manageable steps to utilize the most cost-effective solution for each component.

“

Nobody should be using AI tools just for the sake of using them.