For the first time in a while, the artificial intelligence debate in the markets entered a new phase. The usual narrative is that AI will mostly boost software companies but in the last few days, the script is changing.
Broadly, cracks appeared in the narrative across three segments: software, data analytics, and professional services stocks.
The reason? Anthropic rolled out new plug-ins for its Claude Cowork agent. Anthropic’s move has major implications across legal research, data analysis, sales, and marketing.
Markets didn’t see “incremental productivity.” Instead, they saw business model risk.
The AI honeymoon ends for software as investors rush for the exits
European and U.S. software stocks are on the back foot. They are struggling to stabilize as the sell-off spreads to cross regions and sectors, pulling in Asia along the way.
Investors are not getting spooked with respect to just the technology. Instead, the speed of the moves is what is terrifying investors.
Related: Bank of America delivers blunt warning on dollars next move
AI agents are moving from demos to real workflows faster than many companies’ pricing models can adapt, and Anthropic’s latest move is no different within this context.
That shift prompted traders to reprice stocks once regarded defensive compounders as potential AI casualties.
Legal and data stocks take the first punch, and it’s a big one
The hardest hits are in areas where AI automation is most on the nose, like legal research, compliance, and analytics.
Shares of RELX and Wolters Kluwer fell sharply as investors pondered whether subscription-based legal and risk analytics can keep their prices high if AI agents can replicate core tasks.
Thomson Reuters (TRI) experienced the most significant impact in the U.S., witnessing its largest decline in years, because of the potential impact to its Westlaw legal research business.
For the longest time, it’s been considered the preeminent profit engine for the company. However, it is now viewed through a disruptive lens. Earnings also did not help, depressing sentiment further.
The market’s rationale is clear. If AI is capable of quickly writing, summarizing, and evaluating legal documents, customers may reconsider their investment in conventional tools.
This is the Anthropic move that flipped the AI narrative overnight
Whenever new products debut, a common perception is that they are just flashy consumer features. However, it’s a mistake to treat Anthropic’s Claude Cowork as nothing more than a showy plug-in.
Such features are enterprise workflow tools.
The plug-ins allow businesses to:
- Define standardized workflows for AI agents
- Pull from internal and external data sources
- Automate repeatable knowledge-intensive tasks
Who are the casualties of Anthropic’s bold vision, you might ask? The companies that charge per seat, per query, or per user are now under pressure.
The markets do not want to see how Anthropic’s move will play out. Instead, it’s sentencing first and asking questions later.
Salesforce and Adobe wobble as AI anxiety spreads beyond legal tech
Legal and data providers are not the only ones facing challenges.
Large mega-tech software companies such as Salesforce (CRM) and Adobe (ADBE) are losing value because investors are worried that generative AI could lower prices for CRM, marketing, and creative tools.
Related: SoFi CEO drops a bombshell after blowout earnings
The stocks are not collapsing, but the tone is different now. Instead of believing that AI is pure upside, investors are wondering what benefits consumers may eventually get for free from AI agents.
Even CrowdStrike (CRWD), a cybersecurity company, felt the strain as tech beta spread to all areas.
If AI can do the work, what happens to agency fees?
Advertising stocks also felt the pinch. Shares of Publicis and WPP fell as investors thought about an issue that made them uncomfortable: How defensible are agency fees if AI can handle creative work, targeting, and optimization?
It wasn’t about missing profits. It was about being relevant for a long time.
When AI dread shifts from legal research to ad creative, it means people need to think more about how many “knowledge services” still need big teams of people.
Why AI winners and losers are starting to separate
Not every tech stock is under the pump. Infrastructure and platform players including Nvidia (NVDA) and Microsoft (MSFT) are performing much better. The sharp contrast in fortunes is thanks to reinforcement of a growing divide in the AI trade.
The market appears to favor companies that:
- Sell the picks and shovels (chips, cloud, compute)
- Control distribution at scale
- Can bundle AI into must-have platforms
Meanwhile, the enterprises employing standalone software tools, they are the ones in trouble.
The 3 questions every software investor now has to answer
For regular investors, the selloff brings up several important considerations as earnings season approaches:
- Is the data from this firm really private, or can AI readily copy it?
- Can it move away from pricing models based on the number of seats?
- Does AI make it easier or harder for customers to stay with a company?
Stocks that answer those questions well might bounce back rapidly. Some people may have to wait longer for a more substantial reset.
There is no AI panic, just repricing
The selloff isn’t a verdict on AI. Instead, it’s a verdict on who captures the value.
For my money, AI is still creating enormous opportunity, but the share of the pie is not being distributed evenly.
For investors, the era of treating “software” as a single, safe bucket is ending; Anthropic has shown that much. From here, business models matter more than buzzwords.