Claude Cowork’s Blitz: AI Agent Ignites Software Sector Selloff

In a stark demonstration of artificial intelligence’s disruptive force, Anthropic’s launch of Claude Cowork has triggered a sharp downturn in software stocks, sending the iShares Expanded Tech-Software Sector ETF into bear-market territory. Shares of ServiceNow plunged nearly 11% on January 29, 2026, despite beating earnings expectations, as investors grappled with fears that autonomous AI agents could erode demand for traditional enterprise software. The tool, unveiled on January 12 as a research preview, enables Claude to directly manipulate files, organize desktops, and execute workflows on users’ computers, raising questions about the viability of subscription-based models that have underpinned the sector for years. CNBC reported the ETF’s 5.4% single-day drop, its steepest since last April, with the fund now down 22% from recent highs.

Microsoft shares tumbled 10%, shedding $357 billion in market value in their worst day since March 2020, even as cloud growth slowed less than feared. SAP fell 15.2% after cloud backlog growth missed estimates at 16%. HubSpot, Klaviyo, and Atlassian each lost over 10%, reflecting broad unease. Ben Reitzes, head of technology research at Melius Research, captured the sentiment: “It is a little embarrassing that in 10 days, Anthropic was able to invent, co-work, put it out and everybody … could look at it and go, ‘Wow, why isn’t Microsoft doing that?’” Yahoo Finance highlighted how Claude Cowork’s viral success exacerbated a 15% year-to-date decline in SaaS stocks.

ServiceNow CEO Bill McDermott pushed back, emphasizing the need for reliable workflow software amid AI’s probabilistic nature: “The real payoff comes when trillions of tokens move beyond pilots to be embedded directly into the workflows where business decisions are made. ServiceNow is the gateway to this shift.” Yet, Morgan Stanley analysts deemed results “good, but not good enough” in an environment of heightened skepticism. The episode underscores a pivotal tension: incumbents tout AI integrations, but investors fear general-purpose agents like Cowork could bypass specialized tools entirely.

Anthropic’s Rapid AI-Built Breakthrough

Boris Cherny, head of Claude Code at Anthropic, revealed that Cowork was developed in about 10 days, with Claude Code authoring the vast majority of its codebase. “We might be six to 12 months away from when the model is doing most, maybe all of what software engineers do end-to-end,” CEO Dario Amodei predicted. This self-reinforcing development cycle—AI building AI—marks a leap in efficiency, positioning Anthropic at $350 billion valuation amid a $10 billion fundraising round, per Yahoo Finance .

Cowork extends Claude Code’s terminal-based powers to non-technical users via a macOS desktop app, initially for $100-$200/month Max subscribers, later expanded to $20/month Pro tiers. Users grant scoped access to folders, where Claude reads, edits, creates files, organizes screenshots into spreadsheets, drafts reports from notes, and browses via a Chrome extension—all from plain English prompts. It runs in a sandboxed virtual machine with deletion protection, addressing risks like unintended file removal or prompt injections. WIRED tested it organizing desktops and booking events, noting reliable performance for a beta.

Anthropic’s blog and executives frame Cowork as “Claude Code for the rest of your work,” born from users repurposing the coding tool for tasks like vacation planning or photo recovery. Rahul Patil, CTO, noted on LinkedIn: “When we released Claude Code, we expected developers to use it for coding. They did—and then quickly began using it for almost everything else.” This evolution targets knowledge workers—analysts, marketers, accountants—democratizing agentic AI.

From Chatbot to Desktop Dominator

Unlike chatbots confined to text, Cowork employs visual reasoning to navigate interfaces, move cursors, and type like a human, using Agent Skills for formats like XLSX, PPTX, DOCX, and PDFs. It spawns sub-agents for parallel tasks, integrates with connectors like Google Calendar or AWS, and leverages the Model Context Protocol for interoperability—adopted by Microsoft VS Code and OpenAI Codex. Ars Technica described it bringing Claude Code’s modality to office work without technical setup.

Simon Willison’s hands-on review confirmed sandboxing via containerized mounting, ensuring access limits. DataCamp tutorials showcased expense reports from receipt photos and multi-source reports. Fortune warned it threatens startups in file management, while InfoQ praised sub-agent coordination. Limitations persist: macOS-only initially, no session memory, slower browser automation, and xlsx parsing issues, but trajectory points to Windows support and refinements.

Security layers include prompt-injection detection and user approvals for actions, though Anthropic cautions against sensitive data exposure. “Agent safety… is still an active area of development,” the company stated. Early adopters hail it as transformative; one engineer told WIRED : “This is the most fun I’ve ever had… I don’t have to do all the tedious work.”

Wall Street’s AI Disruption Reckoning

The iShares ETF, down over 13% month-to-date, nears its worst month since 2008. Mizuho’s Jordan Klein noted buy-siders see “no reasons to own software,” per Bloomberg via Sherwood News. RBC analysts flagged “AI bear narrative contagion” spreading to verticals like healthcare after Claude’s specialized launches, warning of 2026’s innovation velocity. Sherwood News tied it to only 31% of holdings above 200-day averages.

Atlassian dropped 16% post-launch, per Australian Financial Review ; Salesforce and Adobe saw modest declines. C3.ai faces sentiment pressure, with William Blair’s Arjun Bhatia calling reactions “overstated” but acknowledging agentic threats. ServiceNow deepened its Anthropic partnership, embedding Claude models, yet Wedbush’s Dan Ives removed it from top AI picks over slow monetization.

Paradoxically, Cloudflare rose on derivative bets for AI workload infrastructure like Cowork and Clawdbot. Anthropic’s $1 billion ARR from Claude Code, eyeing $1.1 billion, fuels its edge, with Claude reportedly outpacing ChatGPT in enterprise. As agents like OpenAI’s Operator compete, 2026 emerges as the “Agent Wars” year, per analysts, challenging SaaS’s asset-light, high-margin premise with outcome-based “Service-as-Software.”

Incumbents Scramble, Agents Ascend

Microsoft directs engineers to Claude Code despite OpenAI ties, investing up to $5 billion in Anthropic. Executives like Amodei foresee AI handling end-to-end engineering soon. Business Insider linked selloffs in Salesforce, Workday to announcements. Investors weigh disruption against incumbents’ moats—data, integrations, reliability.

Cowork signals a paradigm where value shifts from tools to outcomes: organized folders, completed reports. Amazon, an investor, eyes agentic futures in cloud and logistics. Early Reddit buzz and Wired interviews reveal “Claude-pilled” engineers five times more productive. As Anthropic dogfoods its tools—95% of Claude Code team code via AI—it accelerates toward profitability by 2028, projecting positive cash flow in 2027.

For software giants, the pressure mounts: adapt to agentic reality or risk obsolescence. Claude Cowork isn’t just a tool; it’s a harbinger of AI’s migration from advisor to executor, forcing a reevaluation of trillion-dollar valuations built on predictable subscriptions now vulnerable to autonomous alternatives.