If you have ever winced at your monthly AI bill, here’s some good news. According to a report by The Wall Street Journal, OpenAI is considering drastically lowering the prices it charges users as it fights to win customers from its rival, Anthropic.

The company is weighing significant cuts to its token pricing, the unit AI firms use to bill for their products. Interestingly, the move is in anticipation of similar cuts OpenAI expects from Anthropic. So whichever AI service you use, your bills should get smaller.

Why is OpenAI suddenly feeling generous?

The answer is simple: businesses are tired of paying sky-high prices for AI. Heck, there have even been reports of AI costing companies more than actual employees. Even OpenAI CEO Sam Altman admitted at a recent event that costs had become “a huge issue,” adding, “I think we’ll have a lot of ways we can help people get more value for less spend.”

But that’s not all. OpenAI is also facing stiff competition. Anthropic’s revenue surged after its coding tool, Claude Code, went viral among software engineers, and the five-year-old startup surpassed OpenAI’s valuation for the first time. OpenAI has since made its own coding tool, Codex, a company focus, but it’s still far behind the competition.

Some corporations poured so much money into AI coding tools that their leaders are now reining in spending. An Uber executive said the company had already maxed out its 2026 budget for agentic AI. These comments have sparked a Silicon Valley debate about tokenmaxxing, the practice of burning through as many tokens as possible to boost productivity, even when it doesn’t generate returns.

Then there’s Google. Its Gemini models, especially the budget Flash tiers, undercut both ChatGPT and Claude on price, and its business plans cost nearly half of what OpenAI charges, adding even more competitive pressure.

What does a price war mean for you?

For the companies, it’s risky. Both companies already lose billions on computing costs, and both have confidentially filed for IPOs. Slashing prices right before facing public investors will be the first real test of their business models.

For users, it’s good news. They will soon see a drastic reduction in their AI costs. Competition is always good for consumers, and a price cut is one of the big benefits. So sit back and let the AI giants fight it out, because for once, we are the ones who win.

TL;DR

Enterprise AI bills are tripling despite a 98% drop in per-token prices, as agentic tools drive consumption 18.6x higher per developer. The Linux Foundation is launching the Tokenomics Foundation to bring cost discipline to AI spending.

Uber blew through its entire 2026 AI coding budget by April. Microsoft revoked its developers’ Claude Code licences six months after enabling them. One company reportedly ran up a $500 million Claude bill in a single month after forgetting to set usage limits. A Priceline employee told TechCrunch that a routine Cursor contract renewal came back four to five times more expensive.

The pattern is the same everywhere. Per-token prices have collapsed, but the push for autonomous AI agents has sent consumption through the roof. Companies that gorged themselves on all-you-can-eat subscriptions in early 2025 are now scrambling to understand where the money went, and whether any of it produced a return.

The paradox in numbers

GPT-4-equivalent performance now costs roughly $0.40 per million tokens, down from $20 per million in late 2022. That is a 98% reduction. Yet enterprise AI bills have risen by an estimated 320%, according to multiple industry analyses. The average enterprise AI budget has grown from $1.2 million per year in 2024 to $7 million in 2026.

The culprit is volume. Agentic AI tools released since November 2025, including Anthropic’s Claude Opus 4.5, OpenAI’s GPT-5.1, and Google’s Gemini 3 Pro, have multiplied token consumption per task. A simple linear workflow in 2023 cost about $0.04 per interaction. An orchestrated agentic system in 2026 costs roughly $1.20, about 30 times more. Individual engineers at Microsoft were reportedly spending between $500 and $2,000 a month on tokens before the licences were pulled.

Nicholas Arcolano, head of research at engineering management platform Jellyfish, told TechCrunch that per-developer consumption has risen roughly 18.6 times in nine months. Engineers who used the most tokens were about twice as productive as lighter users, but they spent 10 times the tokens to get there. “Whether extreme spend pays off comes down to the ultimate business value of shipped code, which most companies still can’t measure,” Arcolano said.

From tokenmaxxing to guardrails

“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?’” Alexander Embiricos, OpenAI’s head of enterprise, told TechCrunch. “Now the conversations are about, ‘We’re spending so much. What visibility do you have? What token controls do you have?’”

J.R. Storment, executive director of the FinOps Foundation, described the shift bluntly. “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.’ The whole conversation shifted from tokenmaxxing and ‘go fast’ to ‘we need guardrails, how do we control this?’”

Priceline’s senior director of IT finance, Chris Reed, drew a comparison to the telecom billing era. “It’s like the crack-cocaine epidemic. They let you try it to get you hooked, and now you’re kind of beholden to it.” The company has begun placing token limits on certain groups. Reed said he is already seeing discrepancies between vendor-reported usage and Priceline’s internal data.

The Tokenomics Foundation

It is against this backdrop that the Linux Foundation this week unveiled plans for the Tokenomics Foundation, a new standards body aiming to bring the same cost discipline to AI tokens that FinOps brought to cloud spending.

The Foundation plans to build a canonical definition of “tokenomics,” open standards for AI token usage and billing, and new metrics including cost-per-intelligence and tokens-per-watt. A formal launch is planned for July. Nishant Gupta, chief availability officer at Salesforce, said in a statement that “token economics is fundamentally more abstract and opaque than anything we’ve managed at this scale before.”

The challenge is enormous. “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.”

A market forms around the problem

Startups and established vendors are racing to fill the gap. Pay-i tracks and optimises AI spending. Paid lets developers bill based on actual value rather than subscription fees. Jellyfish, Waydev, and Faros AI provide agent monitoring to prove the ROI of developer tools. Ramp has moved into AI spend management. Datadog and New Relic have added token-level observability.

Model routing is emerging as the primary cost lever. Factory, an enterprise AI coding startup, launched a model router this week that automatically picks the cheapest adequate model for each task. Vitaly Gordon, CEO of Faros AI, said frontier labs are already doing this internally. “The financial report for how much you spend on Anthropic, even if you call the Opus model, some of the spend will be on Sonnet or Haiku, because they are smart enough to do it,” he said.

Goldman Sachs projects global token usage will multiply 24 times by 2030. The companies already over budget need solutions now, and the Tokenomics Foundation’s first deliverable is still months away. As Gordon put it: “Maybe we created a steam engine, but we still haven’t figured out the assembly line.”

CBA chief executive Matt Comyn used the phrase ‘work slop’ to describe the low-quality AI output now flowing through corporate workflows, as token-billed AI costs scale with task complexity.

Matt Comyn, chief executive of the Commonwealth Bank of Australia, used a speech on Monday to flag two AI-adoption problems large corporate buyers have been working through quietly for several months.

The first is that the cost of running generative AI inside corporate workflows is rising substantially faster than most companies budgeted for as task complexity scales.

The second is what Comyn called “work slop”, the low-quality AI-generated text, code and analysis that flows through internal company systems when employees use AI without sufficient quality control.

The cost framing is the part that will resonate with the corporate-IT-buyer audience. Token-based pricing, the per-character billing model the foundation-model labs use to charge enterprise customers, has scaled in the past 18 months from a modest line item into a meaningful operating-expense category.

Comyn’s point is that the cost compounds faster than expected because token consumption per task rises non-linearly with task complexity: a simple summarisation task might consume 1,000 tokens, but a multi-step reasoning task with tool use can consume 100,000-plus tokens for the same output value. Companies that priced their AI rollouts on the simple-task baseline are now seeing bills that scale on the complex-task curve.

This problem is not specific to CBA. Morgan Stanley doubled its European-banking-AI-job-loss forecast last week partly on evidence that AI cost-benefit ratios are tightening at exactly the moment large institutions had hoped they would loosen. The token-cost-scaling problem Comyn described is the underlying mechanic: the same AI deployment that worked at pilot-stage volumes can produce 10-100x the costs at production-stage complexity.

The result is the corporate-AI procurement squeeze that Comyn predicted will tighten through 2026: businesses tightening scrutiny of AI-related spending as pressure mounts to demonstrate returns on investment.

The “work slop” framing is the more colourful but equally substantive half of the speech. The category Comyn was naming, low-quality AI-generated output that nominally completes a task but actually degrades downstream workflow, is the corporate-knowledge-work analogue of the social-media “AI slop” problem that emerged in 2024 with image-generation tools.

The bank version looks like this: an employee uses ChatGPT to draft a customer email, the email is technically grammatical but factually imprecise, the recipient takes the imprecision as a commitment, and the bank deals with the resulting complaint three weeks later at a substantially higher cost than the original work would have generated unaided.

The CBA-specific context is significant. The bank announced 90 job cuts earlier this year and a further 120 cuts in May explicitly attributed to AI-driven productivity gains, alongside a A$90m AI-workforce reskilling commitment.

Comyn’s remarks therefore land inside a CBA strategy that has visibly committed to AI substitution at scale: the “work slop” framing is not a defensive critique of AI from a bank that has rejected the technology but a sharper inside-baseball read on AI deployment from one of Australia’s largest current adopters.

The wider Australian-bank context is also worth noting. Sam Altman has been arguing over the past month that an AI jobs apocalypse is unlikely at the macro level, and the labour data through March 2026 has so far supported the conservative read.

Comyn’s remarks complicate that picture: the macro labour data does not yet show large-scale displacement, but the operating-margin data inside large corporates is starting to show the AI-cost-and-quality tradeoffs CBA is now naming explicitly.

The substantive implication is that the 2024-2025 AI cost narrative, that token prices were falling so quickly that the deployment-economics question would solve itself, has structurally inverted.

Falling per-token prices have been overwhelmed by rising per-task token consumption as enterprises move from pilot deployments to production use cases. The procurement-discipline phase Comyn is forecasting through 2026 is, on this evidence, the predictable consequence.