Corporate leaders are starting to question whether soaring AI spending is delivering meaningful returns.

Why it matters: Companies that rushed to embrace AI are now confronting ballooning IT costs, uncertain productivity gains and growing employee skepticism.

Driving the news: Microsoft canceled most of its Claude Code licenses, in part over costs, according to The Verge, and Uber's COO said AI costs are getting "harder to justify."

• An AI consultant tells Axios one of their clients recently spent half a billion dollars in a single month after failing to put usage limits on Claude licenses for employees.
• Companies are citing AI's ability to automate jobs as a cause for layoffs, though Anuj Kapur, CEO of CloudBees, told Axios that workforce cuts may simply be "the only lever they can pull" to offset their AI bills.
• Consumer sentiment around AI is also nosediving, and employees are rebelling against the use of the technology at work.

What they're saying: The enterprise is undergoing a "healthy swing" away from AI overuse — or "tokenmaxxing," the push to burn as many AI tokens as possible — Ali Ansari, CEO of model training firm Micro1, told Axios.

• Ansari hopes this correction will push companies toward more efficient AI use.
• While the market views these tools as working equally well across the enterprise, Ansari says "the reality of AI right now is that it only works for coding."
• That disconnect can drive up IT bills without leading to high return on investment in agents, he said.

Friction point: Corporate AI adoption is running into four unique problems.

• Use cases: "Most people default to automating tasks they dislike rather than tasks most valuable to the company," Sophia Velastegui, CEO of Velastegui Ventures and former chief AI officer at Microsoft, told Axios. Instead, they should focus on using AI to drive revenue.
• Costs: One CTO told Axios that employees were using AI models to check the weather. That gets expensive fast: Enterprise AI plans are not truly "all you can eat," and even simple chatbot queries can carry heavy token costs.
• Humans: We are the bottleneck to more efficient adoption, as we're still catching up on AI. Leadership isn't always helping: Throwing AI licenses at the wall and seeing what sticks (or what Velastegui calls the "thousand flowers bloom" approach) isn't leading to tangible returns, she said.
• Data: When enterprises are hesitant to give AI agents unfettered access to proprietary data, those agents become less effective, Josh Pantony, CEO of Boosted.ai, which focuses on AI tools for finance, told Axios.

What we're watching: Whether companies get more disciplined about AI use. Or overcorrect and clamp down.

Working-age Americans in cities are nearly twice as likely to use AI as those in rural communities, according to a new report from Microsoft, shared first with Axios.

Why it matters: The uneven spread of AI adoption could deepen existing economic opportunity gaps across the U.S., Microsoft president Brad Smith said in an interview.

The big picture: Microsoft's U.S. county-level data shows AI adoption is splitting along familiar economic and geographic lines.

• Part of that boils down to trust. More than half of urban respondents say AI is likely to act in the public interest, compared with less than 40% in rural areas.

By the numbers: Nearly a third of people in large urban areas use AI, compared with 16.2% of rural residents.

• Residents of smaller cities fall in between those extremes, with about 22% of people using AI.
• There's some variance even within large cities. Among the 35 largest U.S. metro areas, AI usage ranges from a high of almost 40% in Washington, D.C., to just over 25% in Pittsburgh.

Between the lines: Smith said that closing the rural-urban gap is important — and not just to tech companies that want more customers.

• Deploying AI more broadly, Smith hopes, will generate economic growth across the country.
• "It's very unfortunate if the people who could benefit from it the most — who arguably need it the most — are accessing generative AI less frequently," he said.

Friction point: The products themselves are part of the problem, Smith admits.

• "We have to understand people's problems. We have to make the case and we have to make the products useful and easy for people to use," he told Axios.
• "We have a lot of work to do," he said.

Zoom out: Smith says the report should be a wake-up call and not just for those using — or not using — the technology.

• "I think that it is an imperative for the tech sector not only to make the case, but to heed the importance of building AI in a way that gives people the opportunity to pursue better jobs," Smith said.
• "Usually when you have a PR problem, it's because you have a reality problem."

Reality check: College towns are hotspots of AI usage.

• Every county in the top 15 AI adopters is home to a college or university.
• Smith noted that while college-aged people are the heaviest users of AI, they are also among the technology's loudest critics.

The bottom line: Closing the AI adoption gap could determine whether the technology narrows or widens economic divides across the country.

OpenAI is announcing new partnerships to combat misinformation, offering its cybersecurity products to state officials and backing legislation ahead of elections in the U.S. and globally.

Why it matters: AI is becoming an election mainstay, with candidates, especially Republicans, using it for their campaigns as voters turn to chatbots for information.

Driving the news: OpenAI is seeking to meet the moment with a series of new efforts shared first with Axios yesterday.

• The company is offering its cybersecurity products — Codex Security and its Trusted Access for Cyber program — to registered voting system manufacturers in the U.S.
• It's briefing the National Association of Secretaries of State and the National Association of State Election Directors on the latest cyber capabilities.
• OpenAI will provide live vote counts from the Associated Press beginning this fall in the U.S. and Brazil, and is partnering with Democracy Works to display reliable information about voting and registration processes.

The company is also endorsing transparency legislation to combat deepfakes, including the Protect Elections from Deceptive AI Act and Preparing Election Administrators for AI Act.

The big picture: AI companies are getting around to what social media companies have had to reckon with since 2016 — their tools have the power to influence elections.

• OpenAI made similar efforts ahead of the 2024 election, during which AI companies faced widespread criticism for inaccurate and misleading information being generated by chatbots.
• In just two years, the technology has improved exponentially, and voters are set to face a whole new level of AI-generated upheaval.

• Chipmaker Groq is raising $650 million after delivering a major win to its early investors, Dan Primack reports. (Axios)
• Meta is testing paid AI subscription plans, starting at $7.99 per month, while CEO Mark Zuckerberg told shareholders the company can always rent out excess compute power to businesses if it winds up overbuilding. (CNBC)
• A federal complaint accuses a Google engineer of using confidential company data to make $1.2 million from placing bets on Polymarket. (Axios)

A Ford dealer in Kansas has a truck he really can't sell.