The rush to adopt artificial intelligence is creating a new challenge for businesses: controlling the cost.
Across the industry, companies are finding that AI usage is rising far faster than expected. The surge has forced many to rethink budgets, impose limits and search for ways to prove returns on their investments.
The conversation around AI has changed markedly in recent months. For many businesses, the question is no longer whether the technology works, but how much it costs to run.
“Six months ago, I would have a conversation with a customer and it would be all about ‘What can it do? Is it good enough?’” said Alexander Embiricos, OpenAI’s head of enterprise.
“Our conversations are never about that now. Now the conversations are about, ‘hey, we’re spending so much. What visibility do you have? What auditability do you have? What token controls do you have? What is the efficiency of your models?’”
Spending is blowing past budgets
Several companies have already encountered problems.
Among the examples cited:
- Uber exhausted its 2026 AI coding budget by April.
- Microsoft withdrew developers’ Claude Code licenses months after introducing them.
- A Priceline employee said a routine Cursor contract renewal became four to five times more expensive.
Falling per-token prices have done little to curb overall spending. Wider adoption and increasingly autonomous AI agents have pushed usage sharply higher.
J.R. Storment, executive director of the FinOps Foundation, said companies began raising alarms earlier this year.
“In April and May, I started hearing from companies: ‘Oh my god, we are 3x over our entire 2026 token budget and it’s only April,’” he said.
“We started hearing existential crises, and the whole conversation shifted from tokenmaxxing and ‘go fast’ to ‘we need guardrails, how do we control this?’”
New models are driving higher consumption
Executives say more advanced AI systems have accelerated spending.
Among the models released in November were:
- Anthropic’s Claude Opus 4.5
- OpenAI’s GPT-5.1
- Google’s Gemini 3 Pro
Their improved capabilities have fueled the use of AI agents, which consume far more tokens.
One company reportedly ended up with a $500 million Claude bill after failing to impose usage limits.
“It’s like the crack-cocaine epidemic,” said Chris Reed, senior director of IT finance at Priceline.
“They let you try it to get you hooked on it, and now you’re kind of beholden to it.”
Productivity gains do not always justify the expense
The return on AI spending remains difficult to measure.
Vitaly Gordon, chief executive of Faros AI, recalled a conversation with a technology executive wrestling with the issue.
“One of my engineers spent $40,000 on tokens last month, and I genuinely don’t know whether I should stop him or should I go and tell everyone else to be like him,” the executive said.
A two-year Faros study involving 20,000 developers found output increased, but so did bugs and code rewrites.
Research by Jellyfish reached similar conclusions. Developers who used AI heavily were roughly twice as productive as lighter users, but consumed around ten times as many tokens.
“Whether extreme spend pays off comes down to the ultimate business value of shipped code, which most companies still can’t measure,” said Nicholas Arcolano, head of research at Jellyfish.
Tracking AI costs is becoming harder
The scale of AI usage is creating new challenges for finance teams.
“Tracking cloud costs is a hundreds-of-millions-of-rows-a-month data problem,” Storment said.
“Tracking token costs is a trillions-of-rows-a-month data problem.”
Chris Reed said the industry is already encountering discrepancies between vendor reports and internal records.
“Anytime you introduce something new, it’s ripe for billing errors and audit and optimization opportunities,” he said.
New market emerges around AI spending
A growing number of companies are developing tools to monitor AI usage and costs.
Specialized firms include:
- Pay-i, which tracks and optimizes generative AI spending.
- Paid, which helps developers measure usage and bill customers.
Engineering platforms such as:
- Jellyfish
- Waydev
- Faros AI
are also adding AI monitoring capabilities.
Established players are expanding into the sector as well. Recent additions include:
- Ramp
- Datadog
- New Relic
AWS is expected to unveil new enterprise AI financial management tools at the FinOps X conference next week.
Industry seeks common standards
The Linux Foundation this week unveiled plans for the Tokenomics Foundation, a standards body aimed at bringing more discipline to AI spending.
The group plans to create:
- Standard definitions for token usage and billing.
- Common metrics for comparing costs across vendors.
- New measures such as cost-per-intelligence and tokens-per-watt.
- Metrics covering consumption efficiency.
A formal launch is planned for July.
“Token economics is fundamentally more abstract and opaque than anything we’ve managed at this scale before,” said Nishant Gupta, chief availability officer at Salesforce.
“It requires a different operational muscle than the one the industry built for cloud.”
Demand is expected to keep rising
Goldman Sachs projects global token usage will increase 24-fold by 2030.
For many companies, however, the need for solutions is immediate.
“Maybe we created a steam engine, but we still haven’t figured out the assembly line,” Gordon said.
Arcolano said businesses may get better returns by spreading AI adoption more broadly rather than concentrating usage among a handful of heavy users.
“The best ROI comes from moving the broad middle from low to moderate usage, not pushing heavy users higher,” he said.
TL;DR:
Companies are struggling to control exploding AI costs as token usage rises faster than budgets. New vendors and the Linux Foundation’s Tokenomics Foundation are working on tools and standards to help enterprises measure spending and improve returns.
AI summary:
- Companies are exceeding AI budgets as token consumption surges.
- More powerful AI agents are driving much higher usage.
- Measuring returns remains difficult despite productivity gains.
- New vendors are building tools to monitor AI spending.
- The Linux Foundation plans to launch the Tokenomics Foundation in July.
