The $500 Million Wake-Up Call

Axios reported this week that an Anthropic enterprise client — identity undisclosed — spent roughly $500 million on Claude in a single month. The reason: no usage limits on employee licenses.

Half a billion dollars. One month. One tool. Zero guardrails.

ZeroHedge noted the timing was striking — the Axios report dropped just as Amazon was shutting down an internal AI-usage leaderboard after employees started 'tokenmaxxing' — deliberately routing unnecessary work through AI tools to inflate their adoption scores.

Amazon turned AI adoption into a metric. Employees optimized for the metric instead of the work.

Uber, Microsoft, and the AI Hangover

Amazon isn't alone. According to ZeroHedge, Uber burned through its entire 2026 AI coding-tools budget by April — eight months early. COO Andrew Macdonald admitted the company found it "very hard to draw a line" between real productivity and AI noise.

Microsoft has reportedly begun canceling most Claude Code licenses and pushing developers toward GitHub Copilot CLI instead. This from the company that has been OpenAI's most visible cheerleader.

These aren't startups learning on the fly. These are Fortune 500 operations with armies of finance staff. And they got burned.

The Fundamental Math Problem CFOs Ignored

Glean CEO Arvind Jain told CNBC bluntly: companies are burning through annual AI budgets in one to two months. He called it "an unsustainable path right now."

Each new frontier model release costs roughly twice as much per token as the model it replaced. That means the more powerful AI gets, the more expensive deployment becomes — even as companies are being sold the narrative that AI will drive efficiency savings.

Factory AI CEO Matan Grinberg told CNBC that companies have now landed on a single defining question: optimize the number of employees, or optimize AI spend per employee?

Roughly 95% of enterprise AI is still running on the most expensive frontier models — even for tasks that could run on cheaper, smaller models. According to CNBC, routing simpler work to cheaper model tiers could deliver significant savings. But companies aren't doing it. They bought the Ferrari and are using it to drive to the grocery store.

The Layoffs Are Already Here

While the budget chaos plays out in finance departments, the human cost is arriving in real time.

According to Business Insider's updated tracker, at least 14 major companies have now explicitly cited AI efficiencies as a reason for cuts:

• Atlassian cut 1,600 jobs — roughly 10% of its global workforce — in March, citing AI restructuring
• Angi eliminated 350 jobs in January, citing "AI-driven efficiency improvements"
• Block (Jack Dorsey), Salesforce (Marc Benioff), and IBM (Arvind Krishna) are all on the list

A March 2026 report from career transition firm Challenger, Gray, and Christmas found AI has been cited in 8% of all job-cut plans so far this year. That number is growing.

Tech.co's tracking adds Microsoft to the picture — around 6,000 employees cut in May 2025, predominantly programmers, following CEO Satya Nadella's admission that roughly 30% of the company's code is now written by AI.

What Mainstream Media Is Getting Wrong

Left-leaning outlets are covering the layoffs through a worker-protection frame — real concern, but they're largely ignoring the budget implosion angle that explains WHY the layoffs are accelerating right now.

Right-leaning coverage, including ZeroHedge's solid reporting on the tokenmaxxing fiasco, correctly identifies the corporate governance failure but undersells the structural cost problem — this isn't just companies being reckless, it's that the underlying economics of frontier AI are genuinely broken for enterprise use at current pricing.

The Atlantic's take zooms out to historical parallels with the Industrial Revolution — thoughtful, but too abstract. Regular people don't need a history lesson. They need to know their job is on the list.

Three threads have yet to be connected simultaneously: ballooning costs, layoffs justified partly by those costs, and AI tools that aren't even delivering the ROI promised. An MIT study cited by Business Insider found 95% of corporate AI investments have generated zero return so far. Meanwhile Sam Altman himself told reporters some companies are using AI as cover for layoffs that would have happened regardless.

What This Means for Regular People

If you work in knowledge work — coding, customer support, marketing, HR, finance — your employer is actively running the math on whether tokens are cheaper than you.

Some will get that math wrong and blow their budgets like the unnamed Anthropic client. Others will get it right and cut headcount.

A 2025 Robert Half survey found 29% of hiring managers reopened positions after AI implementations underperformed. Some jobs come back. But "some come back" is cold comfort when you're the one who got cut first.

The AI bill is here. Corporate America is figuring out who pays it — and right now, it's workers and shareholders sharing the tab.