Will we still need software in the future, or will AI be enough to handle all kinds of tasks on computers? The latter possibility is becoming increasingly real. Most recently, with the release of Claude Cowork by AI giant Anthropic. Anthropic has introduced Cowork as an AI tool that is directly connected to the best LLM in the Claude series. It’s an AI tool that not only lets you write and create programs, but also takes over commonly tedious computer tasks, from analyzing data to drafting legal documents to preparing for meetings.
And that’s now sending shockwaves through the stock market, especially those of SaaS companies that make money from work. Claude Cowork was only announced on January 12th, but its impact is already visible. While Wall Street is trading near all-time highs, even Microsoft is suffering losses.
The company, long the world’s most valuable company, is down 9% since the start of 2026 and trades 21% below its yearly high of $555. The ranking of the most valuable companies has changed. Microsoft is currently only in fourth place behind Nvidia, Google, and Apple. At the same time, major U.S. indexes fell sharply, with the tech-heavy Nasdaq Composite down 1.4% and the broader S&P 500 down 0.8%.
The trigger for this sudden drop was the product announcement by AI company Anthropic. The company announced Claude Cowork, a platform that provides AI agents for specific business tasks, including tools for legal operations that automate contract reviews, compliance workflows, and legal briefs.
The market reaction was harsh, with analytics firms like Gartner and S&P Global plummeting 21% and 11%, respectively, and Intuit and Equifax dropping more than 10% each. The JPMorgan index of US software stocks fell 7% in one day, taking its cumulative loss this year to 18%. The entire software aristocracy such as Oracle, Adobe, Salesforce, ServiceNow, DocuSign, Workday, and SAP is rapidly declining. Germany’s most valuable company, SAP, has lost more than a third of its value compared to its highs a year ago.
Business model in crisis
The logic behind inventory loss becomes clear from the structure of the classic Software-as-a-Service model. For many years, revenue was based on user-dependent licenses. More employees mean more workplaces and more revenue. This system worked perfectly until AI systems started replacing, or at least significantly compressing, human work. In the future, if autonomous AI agents take over tasks that previously required entire departments in the enterprise, the question arises: Dozens of Salesforce or Workday licenses will be required?
As AI reviews, drafts, and closes contracts, the need for a DocuSign subscription will decrease. And if a team of graphic designers can generate images and design models that they need in seconds, fewer Adobe licenses are needed. Investors are concerned about explosive growth in productivity and declining revenues for software providers. Markets hate uncertainty. The potential shift from user-dependent models to outcome-based pricing could change the industry’s entire revenue logic. Analysts at Mizuho Securities say many institutional investors currently see no reason to own software stocks, regardless of valuation levels or past losses.
Advertising companies are also under pressure, as Anthropic also offers marketing automation, with Publicis down 9%, WPP nearly 12% and Omnicom down more than 11%. Even exchange operator LSEG, which generates significant revenue from its data and news platform workspace, recorded a 12.8% decline. This is the worst daily performance in five years. Private equity firms that have invested heavily in software in recent years also suffered significant losses. Ares Management and KKR each fell 10%, while Apollo fell 4.8%.
Valuations are at their lowest in 10 years
The current situation is contradictory. Even though software stocks’ price-to-earnings ratios are at 10-year lows, their fundamental metrics remain very strong. Arvey’s fund managers have expressed bewilderment at the disconnect between superior fundamentals and disastrous price performance, a phenomenon rarely seen in this form.
Many investors have already reduced their positions in software stocks in recent months, making them more vulnerable to further selling in new developments. Analysts at Irving Investors report that positioning is at a multi-year low, while risk is at a multi-year high. Jones Trading observers point to the impact of the second round. If AI disrupts the software companies that are customers of hyperscalers like Amazon, Microsoft, and Alphabet, it will also impact the demand for cloud computing power.
The truth probably lies between the two extremes. The software field may be undergoing the biggest period of change since the cloud revolution. AI will not replace software anytime soon, but it may redefine it. Large providers have extensive data sets, deep customer relationships, and critical infrastructure. Companies like Microsoft, SAP, and Salesforce have plenty of revenue outside of traditional software businesses.
SaaS or rather SaaS?
It was already clear in 2024 and 2025 that this software could have a difficult future in the AI era. For example, Klarna loudly announced that it would be discontinuing Salesforce’s CRM cloud and, in the future, also discontinuing Workday, the accounting, human resources, and corporate planning cloud provider (more details here).
A similar statement comes from the world’s largest business software provider. Charles Lamanna, corporate vice president of business applications and platforms at Microsoft, predicted that traditional business applications will be obsolete by 2030. Instead, AI-powered “business agents” are expected to take the lead (more on this here). Now, with the introduction of Cowork, the first effects of this trend are starting to appear in the stock market.
In this context, a new mantra has emerged: “Service as a Software”. Activities that were previously performed by humans as services are being automated and “packaged” by AI software. Instead of hiring accountants, translators, or customer service employees, use AI agents that perform this service independently. The focus is on the results, not the tools. Basically, what you buy is not the tool, but the result of the work.
