A fintech startup called Altruist launched a tool called Hazel on February 10. It reads 1040s, paystubs, trust agreements, and account statements, then generates personalized tax strategies in under three minutes. The same work takes a human financial advisor hours of billable time. It costs $100 a month.
Within hours, LPL Financial dropped 11% intraday before closing down 8.31%. Charles Schwab fell 7.42%. Raymond James lost 8.75%. Stifel followed. By February 18, a second wave hit — Schwab down another 7%, Raymond James and LPL each down 8% again. Morgan Stanley slid 5% over a single week.
Nobody was fired. No regulation changed. No earnings missed. A startup with 848 employees and a $1.9 billion valuation released a product update, and the wealth management sector lost billions in market cap.
The One Percent Problem
The wealth management industry runs on a simple equation: charge 1% of assets under management, justify it with complexity. Tax-loss harvesting, estate planning, Roth conversion ladders — these are the services that make a 1% fee on a $2 million portfolio feel reasonable.
Global AUM sits at $139 trillion. The US alone holds $92.5 trillion in managed wealth. At an average fee approaching 1%, this is a business measured in hundreds of billions of annual revenue. PwC projects average fees will decline from 1.0% to 0.75% by 2030. That projection was published before Hazel existed.
What Hazel does is not new in concept. Robo-advisors have been compressing fees for a decade. What's new is scope. Hazel doesn't just rebalance a portfolio. It reads the same documents a $500-per-hour CPA reads, applies the same tax logic, and produces interactive scenario modeling — what happens if you sell the house, take the bonus, retire early. Three minutes. A hundred dollars.
Morgan Stanley's head of wealth management, Jed Finn, told a UBS conference that AI would be an "accelerant" for their existing 3,500 AI tools. The stock kept falling.
The Scare Trade
Wall Street has a name for what happened: the AI Scare Trade.
CNN reported it might not be over. Bernstein picked eight stocks they think can weather it. BTIG's chief market technician said single-stock moves on AI nerves are "getting more and more extreme" and warned the weakness could drag the broader market.
The pattern started in software. Claude Code and Claude Cowork launched, and investors suddenly reframed per-seat SaaS licensing as a vulnerability instead of a moat. If an AI agent can do the work of a Salesforce seat, why buy the seat?
Then it moved to insurance. A Madrid startup called Tuio built an insurance product on ChatGPT, and insurance broker stocks dropped.
Then wealth management.
Then real estate.
Then trucking.
$3.25 billion moved into energy ETFs in a single month. $1.66 billion left tech. Energy is up 17% year-to-date. Materials up 16.5%. Industrials up 12%. The money is rotating from companies that sell knowledge work to companies that own physical things.
What the Market Is Actually Pricing
This is not a reaction to revenue loss. Altruist manages a fraction of the assets Schwab does. No wealth management firm reported declining income because of Hazel.
What the market is pricing is a category shift. The 1% AUM fee was always a bundle — part investment management, part tax planning, part emotional reassurance, part document handling. AI is unbundling it. The investment management piece was already commodity-priced by Vanguard and Betterment. Now the tax planning piece is commodity-priced too.
What's left is the human relationship — the advisor who talks a client down during a crash, who attends the estate planning meeting with the family, who remembers the grandkids' names. That's real value. But it's not $20,000-a-year-on-a-$2-million-portfolio value.
BlackRock noted that firms acting as intermediaries between data and clients face an existential question. If your primary value is organizing information and applying rules to it, you're competing against something that does it faster, cheaper, and without lunch breaks.
The Deeper Pattern
Every few weeks now, a new sector discovers it's next. Software companies learned in January. Financial services learned in February. The trigger is always small — a product launch, a demo, a feature update — and the reaction is always disproportionate.
That's because the market isn't reacting to individual products. It's reacting to a thesis: agentic AI turns knowledge work into commodity work. The specific app doesn't matter. What matters is the proof of concept.
Altruist didn't have to replace every financial advisor. It had to demonstrate that the hardest part of the job — the part that justified premium pricing — could be automated for the cost of a streaming subscription. The market did the rest.
The rotation into energy and materials isn't irrational nostalgia. It's a bet that companies which own atoms will hold value better than companies which organize bits. Pipelines don't crash 8% because of a product demo. Copper mines don't lose 8% because a startup launched an app.
Whether that bet is right depends on how fast agentic AI moves from tax planning to mineral extraction logistics. Based on the current pace, I wouldn't give it long.
Top comments (0)