Anthropic Claude Cowork winners and losers: Why Anthropic’s Claude Cowork sparked $300 billion tech market sell-off this month – see which sectors are winning and losing amid AI boom
Anthropic’s Claude Cowork raises fears of AI replacing software
The sharp adjustment is concentrated in software, data services and IT outsourcing stocks, and follows the launch of Anthropic’s Claude Cowork, a set of open-source plugins that allow AI agents to autonomously complete tasks from start to finish. In demonstrations, the system independently performed legal research and prepared applications, working with raw inputs rather than traditional software workflows.
That moment changed the way markets viewed artificial intelligence. Rather than viewing it primarily as a productivity tool layered on top of existing software, investors began to view it as a potential replacement for entire categories of software and services. The pace and scale of the sell-off reflected how quickly the change in perception was priced in.
Also read: Spotify shares fall after KeyBanc cuts SPOT stock price target – here’s what investors need to know ahead of tomorrow’s earnings
Anthropic Impact: Salesforce, Adobe and ServiceNow hit as seat-based pricing is questioned
Some of the first forays appeared in armchair-based SaaS companies. Shares of Salesforce, ServiceNow and Adobe fell between 6% and 8% as investors questioned how per-user pricing models would survive if AI intermediaries allowed one person to do the work of many, according to a report.
IT outsourcing firms face pressure as AI automates routine tasks
IT services and outsourcing companies were also hit hard. According to the Forbes report, Indian companies like Infosys and TCS, which rely on billing hours for manual data work and entry-level coding, have seen sharp declines because these tasks are precisely the areas that Claude Cowork’s autonomous plugins can automate.
Legal and data services stocks tumble on autonomous AI concerns
High-margin data and legal service providers experienced even steeper declines. Stocks such as Thomson Reuters and LegalZoom lost 15% to 20% as markets priced in the risk that AI agents could handle legal research, drafting and prioritization.Also read: BTC USD price prediction: Bernstein explains why the latest Bitcoin crash is the weakest bear case in history and predicts the cryptocurrency will reach $150,000 in 2026
Why do proprietary datasets still have long-term value?
At the same time, investors continued to realize that much of the value in these businesses came from proprietary, licensed, and constantly updated datasets that were difficult for large language models to replicate.
Shifting from software subscriptions to outcome-based pricing models
This sale didn’t mean the software itself was gone, but it did underline how business models could change. The traditional per-seat subscription approach looks increasingly vulnerable as AI agents take on more work directly. Newer companies like Palantir and Harvey are already experimenting with results-based billing by charging for tasks completed rather than access to tools, according to a Forbes report.
For users, this could mean the software becomes less visible. Instead of logging into multiple platforms, humans can simply define a goal while the AI agent navigates systems, tools, and data to achieve the outcome.
Nvidia emerges as a key beneficiary of agency AI workloads
While some parts of software have struggled, other areas have emerged as potential winners. Semiconductor companies like Nvidia have attracted attention because autonomous agents require much more computing power than basic text rendering. Reasoning-heavy workloads are driving demand for high-performance accelerators, and Nvidia’s Rubin platform was built with these agency tasks in mind.
Model developers such as Anthropic and OpenAI are also in a strong position. When a single agent can replace multiple software agents by completing tasks end-to-end, the value is concentrated at the model level. Pricing is shifting from access to software to successful outcomes, giving model builders more benefits.
Cloud providers Amazon and Google leverage always-on AI agents
Cloud providers including Amazon and Google will also benefit. Autonomous agents operate continuously rather than sporadically, placing higher demands on compute, storage, and power. In this environment, the companies that control data centers, energy access, and infrastructure are starting to look like the hosts of the always-on agency economy.
Artificial intelligence in the physical world attracts investor attention after software valuation resets
Some investors are also looking beyond pure software. The reset in valuations has revived interest in artificial intelligence in the physical world, where intelligence must be matched with machines operating in the real world. Tesla’s shares have increasingly diverged from traditional software names as it is seen as more of a play on robotics and autonomous mobility through projects such as Optimus and Cybercab.
FAQ
Who benefits from artificial intelligence agents?
Chip makers like Nvidia and cloud providers like Amazon and Google will benefit.
What is Claude Cowork?
It is a set of open-source plugins that allow AI agents to autonomously manage complex tasks from raw inputs.
