Anthropic Open-Sources Claude Cowork Plugins: SaaS "Three-Layer Squeeze" Alert as Wall Street Loses $300 Billion in Market Cap Instantly

News Lead

Recently, AI company Anthropic released 11 open-source Claude Cowork plugins. These plugins are not entirely new models or chatbot upgrades, but practical tool extensions for enterprise workflows. However, this announcement hit like a bombshell, instantly triggering market panic: within just 72 hours, global software, financial services, and asset management stocks lost up to $300 billion in market value. SaaS giants like Salesforce and Workday saw their stock prices plummet, prompting Wall Street to collectively reflect: the rise of AI Agents is quietly disrupting the traditional per-seat SaaS pricing model.

Background: The Low-Key Launch of Claude Cowork Plugins

Anthropic's Claude series models are renowned for their security and enterprise-grade applications. The 11 newly released open-source plugins focus on 'Cowork' functionality, aiming to seamlessly integrate Claude into enterprise workflows. For example, these plugins support tasks like automatic code generation, contract review, and sales follow-ups. Users can configure them simply to allow AI direct access to enterprise data sources such as databases or API interfaces.

Unlike previous model iterations, this release emphasizes open-source and plug-and-play functionality, receiving enthusiastic responses from the developer community. Anthropic officially stated that these plugins aim to 'enhance collaboration efficiency between humans and AI,' but market interpretation far exceeded expectations: they mark AI's transition from 'auxiliary tool' to 'autonomous Agent,' directly threatening the survival of middle-layer SaaS software.

Core Content: SaaS Crisis Under Three-Layer Squeeze

Traditional enterprise software architecture can be divided into three layers: top-layer AI Agents, middle-layer SaaS tools, and bottom-layer data. Imagine a typical company: employees open DingTalk, Feishu, Salesforce and other tools daily for collaboration; management relies on dashboards for decision-making; data is scattered across CRM and ERP systems.

The emergence of AI Agents changes everything. They can directly 'bypass' the fancy interfaces of the middle layer and go straight to the bottom-layer data to complete tasks. Take a sales scenario: previously, salespeople entered customer information in Salesforce, marketing departments used another tool to analyze data, and executives viewed dashboards; now, Claude Cowork plugins can automatically capture customer data, analyze behavior, generate follow-up emails, and even complete transaction closures. Users only need to review the final results.

This 'three-layer squeeze' manifests as: AI Agents above stealing users (automation replacing manual SaaS operations); the data layer below absorbing value (whoever owns high-quality data dominates the ecosystem); middle-layer SaaS being marginalized. The result is obvious: 10 AI Agents can replace 100 employees, and 100 Salesforce customer service agents instantly become zero.

Perspectives: Coexistence of Optimism and Concern

Industry professionals have divided views on this trend. Anthropic CEO Dario Amodei posted on X platform: 'Claude Cowork is not about replacing SaaS, but reshaping collaboration. We believe open-source plugins will accelerate enterprise AI adoption.'

'AI Agents will liberate human creativity, not eliminate employment.' — Dario Amodei, Anthropic CEO

In contrast, from the SaaS camp, Salesforce founder Marc Benioff publicly responded: 'We have deeply integrated AI, like Einstein Copilot. SaaS won't die, it will only evolve.' But analysts' voices are sharper. CB Insights senior analyst Deep Shah noted in a report: 'The growth ceiling of per-head SaaS pricing has appeared. AI-driven 'per-token consumption' or 'pay-per-result' models will dominate the future.'

Wall Street institutions like JPMorgan Chase have also issued warnings: Goldman Sachs reports predict that by 2027, the AI Agent market will reach $500 billion, with traditional SaaS penetration declining by 30%. Chinese companies like Alibaba Cloud and Tencent Enterprise Point executives remain optimistic, believing that local data barriers will protect SaaS transformation.

'The evaporation of SaaS market value is not a bubble burst, but value redistribution. Data and infrastructure are the new kings.' — Deep Shah, CB Insights Analyst

Impact Analysis: Value Transfer and Business Model Reshaping

The money hasn't disappeared, it's just transferred. First and foremost is the pricing model revolution: shifting from '$99/person/month' to AI platform subscriptions (like per-token billing, similar to Alibaba Cloud's model); data becomes the core asset, with platforms like Enterprise WeChat and DingTalk gaining value from accumulated massive interaction data.

Second, security and compliance demands surge. Large-scale AI automation easily amplifies errors, requiring enterprises to pay for monitoring AI behavior and auditing decisions. Traditional audit giants like the Big Four will benefit. Third, pay-per-result models emerge: $5 per contract review, $2 per customer issue resolved—software becomes 'piecework.'

Finally, implementation services explode. While AI tools are numerous, enterprise deployment requires experts to configure workflows, integrate systems, and continuously optimize. This creates new opportunities: entrepreneurs should avoid the 'pretty interface + integration' trap and pivot toward building Agents, data backends, security infrastructure, or deployment services.

For entrepreneurs, if your product is an 'information display' tool, the risk is extremely high—Agents above steal users, data below absorbs value, creating accelerated bidirectional squeeze. Conversely, enterprises focusing on 'automatic task execution' or 'exclusive data' will face the best opportunities in a decade.

Conclusion: Dawn of the New SaaS Era

SaaS isn't dead; the era of simple SaaS has ended. The old model relied on UI and workflow growth; the new model emphasizes data execution and system guarantees. From building 'pretty interfaces' to 'Agents that actually work + data backends + security infrastructure,' the value chain is rapidly reorganizing.

For enterprises building 'real things,' this is transformative dividend. Wall Street's $300 billion evaporation is just the tip of the iceberg, foreshadowing how AI Agents will redraw the trillion-dollar software market landscape. Enterprises need to transform quickly to catch this paradigm shift from 'headcount' to 'results.'