Photographer: Krisztian Bocsi/Bloomberg
(Bloomberg) — For months, investors have been growing increasingly anxious about how artificial intelligence will potentially transform the economy. Last week, those concerns suddenly spilled over into the stock market.
The culprit was AI startup Anthropic, which released new tools designed to automate work tasks in various industries, from legal and data services to financial research. The announcements sparked fears that the innovations would doom countless businesses. In response, investors dumped a broad range of stocks, from Expedia Group Inc. to Salesforce Inc. to London Stock Exchange Group Plc.
Most Read from Bloomberg
By Friday, dip buyers stepped in, helping the widely followed iShares Expanded Tech-Software Sector ETF, better known by its ticker IGV, rebound from its 12% decline over the previous four sessions. But for bleary-eyed Wall Street pros, rattled by days of volatile trading, the message was clear: This is the new reality.
“Things are shipping out weekly, daily,” said Daniel Newman, chief executive officer of the Futurum Group. “The blast radius of companies that could be impacted by AI is going to grow daily.”
Even with the end-of-week rebound, the damage was severe. Thomson Reuters Corp.’s Canada-listed shares plunged 20% on the week, their steepest fall ever. Financial research firm Morningstar Inc. posted its worst week in the stock market since 2009. Software makers HubSpot Inc., Atlassian Corp. and Zscaler Inc. each tumbled more than 16%.
All told, a collection of 164 stocks in the software, financial services and asset management sectors shed $611 billion in market value last week. (Bloomberg LP, the parent of Bloomberg News, competes with LSEG, Thomson Reuters and Morningstar in providing financial data and news. Bloomberg Law sells legal research tools and software.)
AI’s disruptive potential has been a topic of conversation since the debut of OpenAI’s ChatGPT in late 2022. But until last week, most of the attention has been on the winners. With hundreds of billions of dollars being spent to beef up computing capacity, investors eagerly bought the shares of companies considered beneficiaries of the largesse, from chipmakers and networking firms to energy providers and materials producers.
That strategy has paid off handsomely. An index that tracks semiconductor-related stocks has more than tripled since the end of 2022, compared with a 61% advance for IGV and an 81% jump in the S&P 500 Index.
Story Continues
While the so-called pick-and-shovel trade is still winning, the rapid pace of new tools being brought to market by startups like Anthropic and OpenAI, as well as Alphabet Inc.’s Google, is making the long-theorized disruption seem much more imminent. In just the past month, Google roiled video-game stocks with the release of a tool that can create an immersive digital world with simple image or text prompts. And another Anthropic release, a work assistant based on its Claude coding service, sent software stocks tumbling.
The developments added to angst fueled by a set of disappointing earnings reports from software makers late last month. The biggest was Microsoft Corp., which lost $357 billion of market value in a single day after slowing revenue growth in its cloud-computing business fanned anxieties about heavy spending on AI. ServiceNow Inc. sank 10% and SAP SE tumbled 15% following similarly lackluster results.
“It was the stalwarts that failed us,” said Jackson Ader, a software analyst at KeyBanc. “If your results and your guidance aren’t up to snuff, it’s kind of like: What confidence should we have for the rest of the sector?”
While plenty of new names were bruised last week, few have been punished to the extent of traditional software makers, which have been under pressure since last year. Salesforce, which owns the popular team collaboration service Slack, is down 48% from a record high in December 2024. ServiceNow, which makes software for human resources and information technology operations, has dropped 57% since hitting a peak in January 2025.
“I suspect some companies will endure, embrace AI, and prosper, but others will see permanent disruption to their business models or prospects,” said Jim Awad, senior managing director at Clearstead Advisors. “It is very hard to know which is which right now.”
That fear has investors running for the exits. Software is by far the most net-sold group among all sectors since the start of the year, according to Goldman’s prime brokerage desk data. Hedge funds’ net exposure to software hit a record low of less than 3% as of Feb 3, down from a peak of 18% in 2023.
However, there’s little fundamental evidence of deterioration. In fact, in the eyes of Wall Street analysts, the outlook for profits is improving. Earnings for software and services companies in the S&P 500 are projected to rise 19% in 2026, up from projections for 16% growth a few months ago, according to data compiled by Bloomberg Intelligence.
“Everyone is assuming the bottom is going to fall out, in terms of operating metrics. I’m skeptical about that,” said Michael Mullaney, director of global market research at Boston Partners. “It could end up that profits and margins are fine, even if there is disruption. If I were a growth manager, I’d be buying the dip.”
The relentless selling has pushed software stocks deep into territory where technical-minded traders typically expect a rebound. The 14-day relative strength index on the iShares ETF hit 15 on Thursday, the lowest level in almost 15 years, and is around 24 now. Anything below 30 is considered oversold.
Meanwhile, valuations keep getting cheaper. A basket of software stocks tracked by Goldman Sachs sank to a record low of 21 times estimated profits, down from a peak of more than 100 in late 2021, according to data compiled by Bloomberg. Salesforce is trading at 14 times profit expected over the next 12 months, compared with an average of 46 over the past decade.
“We continue to test the valuation floor and then blow right through it,” KeyBanc’s Ader said. “People are gun-shy and skittish to say that these stocks are too cheap, because based on historical multiples you could have made that argument at every point for many many months now, and it wouldn’t have helped you one bit.”
–With assistance from Natalia Kniazhevich.
Most Read from Bloomberg Businessweek
©2026 Bloomberg L.P.