Every major newsroom has an AI reporter right now. Almost all of them came up through tech. Madison Mills came up through Wall Street — interviewing Jamie Dimon and Ray Dalio, building one of the most-read markets newsletters in the country — before Axios made her its senior AI reporter. On The Friday Reporter, Madison and I talked about who's actually funding the AI buildout, what bank executives admit privately but won't say on the record, and whether the AI trade in the stock market is starting to look like the dot-com bubble. Here's what stood out.
Key Takeaways
Madison Mills moved from covering Wall Street to leading Axios's AI coverage because the two beats collided — AI became the biggest story in markets, not just in tech.
Finance executives are telling her privately that fraudsters may currently be winning the AI arms race, even as AI labs share their best models with trusted partners first as a defense.
Hyperscalers like Amazon, Meta and Microsoft are on track to spend nearly a trillion dollars on AI infrastructure this year alone, with no clear point where the spending stops climbing.
Seven of the biggest tech companies now make up nearly 40% of the S&P 500's market cap, a concentration that has investors bracing for a possible correction.
AI's effect on jobs is uneven: entry-level hiring is down roughly 8% this year, even as some companies double down on hiring junior, tech-savvy workers to help them use AI well.
Every company Mills talks to reports the same paradox: employees are busier since adopting AI tools, not less.
Why Is a Markets Reporter Now Covering AI at Axios?
"I like to joke that the markets became so AI-ified that Axios had no choice but to make me an AI reporter," Mills told me, and she's not really joking. She spent years covering the people who move markets, and somewhere along the way, the AI story and the markets story became the same story. She's using that background to press AI labs on the question investors keep asking her: when, if ever, do these companies turn a profit.
That finance-first lens is what separates her coverage from most of the AI beat. She's not just asking what the technology can do. She's asking who's paying for it, who's exposed if it doesn't deliver, and who's being quietly kept in the dark.
Could AI Make Financial Fraud a Bigger Threat to Your Bank Account?
Mills recently reported that AI is raising the risk of large-scale attacks on financial infrastructure and fraud that could drain accounts directly. I asked how close we actually are to that becoming a real crisis. "The conversations I've had with finance executives, particularly off the record and on background, paint a very scary picture about the fraudsters winning right now," she said.
That's part of why AI labs are releasing their most capable models to trusted partners first, she explained — the hope is that "good actors" get the tools to shore up their defenses before bad actors weaponize the same capabilities. There's an upside too, she said: AI is also expanding financial literacy for ordinary people navigating their own portfolios. But the bank sources she talks to are, in her words, "very concerned."
Is the AI Investment Boom Bigger — and Riskier — Than It Looks?
Mills has spent weeks on the phone with Wall Street analysts, and what they'll say off the record is more unsettling than what makes it into print. One source described a small town with a major data center buildout, where a regional bank is quietly funding truckers and construction companies riding the demand wave. "On paper, those aren't going to look like AI loans," Mills said. "But they are."
It's a comparison I'd been testing since a colleague described AI data centers to me as the new fracking, and I put that framing to Mills directly. She ran with it — and pushed it further, into what she calls the shift from a petrodollar economy to a "token dollar" economy, with AI query credits becoming as central to the economy as oil once was. The hyperscalers — Amazon, Meta, Microsoft and their peers — are on pace to spend close to a trillion dollars this year on AI infrastructure alone, and Mills said there's no clear sign of when that spending curve bends back down.
Is There an AI Stock Bubble Right Now?
Seven of the largest tech companies now account for nearly 40% of the S&P 500's total market value, Mills said, and that concentration alone makes a lot of investors nervous. She pointed to the Warren Buffett line that's been on her mind: be greedy when others are fearful, be fearful when others are greedy. "It seems like there's a lot of greed around these tech stocks right now," she said.
She won't call the top. "If I had the answer to whether or not the AI bubble was going to burst, I wouldn't be a reporter — I'd be a millionaire," she said. But she noted the S&P traded at roughly 30 times forward earnings a year ago, a valuation level close to what preceded the dot-com crash, and today's pricing looks somewhat more cautious by comparison — not cheap, just less "priced for perfection."
Does AI Create Jobs or Kill Them?
This is where Mills pushed back on the easy narrative. Public companies have real incentive to tell investors that layoffs are AI-driven, she said, because Wall Street rewards that story. When she's talked to the actual employees at those companies, the AI explanation often doesn't hold up as cleanly as the press release suggests.
Entry-level hiring is a different story — down roughly 8% this year by her count, which tracks with what AI tools can already do. But some employers, especially banks, are going all-in on entry-level hiring anyway, betting that tech-savvy young workers are best positioned to help the whole organization use AI well rather than just fast. Mills sees a pattern she's watched before, in how younger employees drove smart adoption of social media inside companies a decade ago. Senior employees, meanwhile, are defying the stereotype in her reporting — she's finding many of them are the loudest AI evangelists in their companies, not the most resistant, because they have the most to protect.
Is Your Industry Already Feeling AI's Push for "Less Workforce, More Output"?
I told Mills about a conversation with the head of a multinational public affairs firm in Washington, recently downsized and told point-blank that productivity expectations were going up anyway — powered by AI adoption. Mills said that's the pattern across nearly every company she interviews: people are busier since AI tools rolled out, not less busy.
A source at Stanford gave her the comparison that stuck: the washing machine. It was supposed to free up enormous amounts of time. Instead, people simply became more productive with the time they had. Mills thinks AI is following the same arc — not eliminating work so much as raising the bar for how much work fits into a day.
Find Madison Mills
Madison Mills is Axios's senior AI reporter and co-author, with Ina Fried, of the Axios AI+ newsletter, which publishes Monday through Thursday (Axios's policy team covers Fridays). Find her byline at Axios, and follow her on Instagram, TikTok, X and LinkedIn under Madison Mills.
The Friday Reporter is hosted by Lisa Camooso Miller. Listen wherever you get your podcasts — and subscribe so you never miss a conversation.