Hot post on the internet: SaaS is dead, agents should be established

Author: David Ondrej Compiled: Deep Tide TechFlow Deep Tide Introduction: Last Monday, Anthropic released a set of plugins for Claude Cowork. Not a new model, not a chatbot upgrade, but a plugin. Within 24 hours, the software stock market value evaporated by $285 billion. A plugin market announcement erased more wealth in a single day than the annual output of most industries. Wall Street is no longer afraid of AI, they are afraid of what AI will replace. David Ondrej pointed out that what is dying is a very specific type of software business - SaaS built on a weak moat of "beautiful UI+integration+seat based charging". When AI agents can directly complete tasks within existing systems, you don't need 15 SaaS tools with beautiful dashboards. Value is being sucked up into the proxy layer and down into the data layer. Everything in the middle was squeezed. The full text is as follows: Last Monday, Anthropic released a set of plugins for Claude Cowork. It's not a new model, it's not a chatbot upgrade, it's a plugin. Within 24 hours, the software stock market lost $285 billion in value. A plugin market announcement erased more wealth in a single day than the annual output of most industries. Wall Street is no longer afraid of AI. They are afraid of what AI will replace. The following are the areas where most people misunderstand this. They heard 'SaaS is dead' and thought it meant the company would stop purchasing software. It's not like that, not at all. What is dying is a very specific type of software business - if you understand which type, you are looking at the biggest entrepreneurial opportunity in the past decade. Let me explain: In the past 15 years, the SaaS script has been simple: find a common business workflow. Build a beautiful UI around it. Add some integrations. Charge by seat and monthly. Defend your position by converting costs and making minor product adjustments. This script has created hundreds of billionaires. But it has a fatal flaw that nobody talks about. Most of the value never lies in the software itself. It is in the workflow of software organizations. UI is a middleman. And AI has just made intermediaries obsolete. Here's what Anthropic actually did - because the title missed the point. They haven't built better chatbots. They turned Claude into a work execution layer. The Cowork plugin allows AI agents to log into your existing tools - your CRM, your documents, your database - and autonomously execute the entire workflow. Legal audit, sales pipeline management, data analysis, production level code, etc. No manual involvement is required. This is the part that makes the market panic. Because if AI agents can directly complete tasks within your existing system - why do you still need 15 different SaaS tools with beautiful dashboards? This is the part that really should keep SaaS founders up all night: if 10 AI agents can do the work of 100 employees, you no longer need 100 Salesforce seats. AI does not directly kill software. It kills the number of employees who use the software. This kills the revenue model of charging by seat. This killed the business. This is what I call 'Thin Middle Squeeze'. Imagine three layers: top layer - AI agent. Something that actually performs work. Middle layer - SaaS UI. Dashboard, workflow, and the buttons you click. Bottom layer - recording system. Databases, CRM, and ERP that store real data. Now, value is being sucked up into the proxy layer and down into the data layer. Everything in the thin middle layer is crushed. That's why Adobe's forward P/E ratio has dropped from 30 to 12. ServiceNow has decreased from 67 to 28. Not because people don't need what they do - but because investors realize that when AI agents can completely bypass UI, the moat around 'beautiful UI+integration' is as thin as paper. The interface used to be a product, but now it's just a shell. But this means that 'SaaS is dead' and those people are completely wrong. SaaS is not dead, but the easy SaaS moat is dead. This is a huge difference. The company's expenditure on software this year will exceed any previous time. Enterprise AI capital expenditures alone will exceed $470 billion by 2026. This is not a shrinking market - this is an exploding market. Money hasn't disappeared, it's just moving. Most people are too busy with panic to see where it has fallen. Here are the actual places where money flows: AI platform subscriptions are based on usage. Based on consumption volume. Not by seat. The company will pay for AI capacity like paying for cloud computing - based on the content they use, rather than how many people sit in the building. This is already happening. GitHub's AI agents are gated through advanced levels with usage based pricing. This is the template. Recording system agents do not eliminate the backend, they operate on the backend. CRM、ERP、 Data warehouses - these become more valuable, not less important. Because AI agents require clean, authoritative, and trustworthy data to take action. Only garbage data input will generate garbage action output. Large scale companies with standardized data will win. When agents take large-scale actions, errors occur on a large scale in terms of security, governance, and compliance. Every company deploying AI agents will pay for permissions, audit logs, policy execution, monitoring, and evaluation. This is boring infrastructure - it will quietly print money in the next decade. The pricing based on results is no longer '$99 per seat per month', you will see '$5 per review contract'. Each resolved support ticket costs $2. Each rich qualified lead costs $10. Software is priced like labor - because it is replacing labor. This is where the pricing model of the entire industry shifts. The service is surprising. But when building software becomes cheap and easy, companies will try more customized software services. Implementation, workflow design, migration, and integration work - the demand for services is about to explode. Vibe encoding makes creation easy. Running it in real business is a completely different story. So if you are currently building a startup - or considering it - the only important question is: where are you in the stack? If you build in a thin middle layer - building beautiful UI on other people's data, charging by seat, without proprietary advantages - you have serious problems. It's not because your product is not good. But it's because the economics of your location is collapsing in real-time. The proxy layer above you is devouring your interface. The recording system below you is devouring your lock. You are being squeezed from two directions. And this kind of squeezing will only accelerate from here. Below nine is what should be built. Build at the proxy layer. Creating AI native tools that not only display information - they execute workflows. Do not display the dashboard to the user. Complete the work for them. Pricing based on results, not seats. Becoming something of action. Build at the data layer. Having proprietary data. Build a record system for fields that do not yet have a good system. Make yourself the authoritative backend that every AI agent needs to insert. Agents come and go - the data layer is eternal. Building infrastructure. Safety. monitor. evaluate government. Compliance. Enable AI agents to deploy large-scale tools securely. Not sexy. Extremely profitable. The demand hasn't even started yet. Build services. Assist companies in implementing, customizing, and operating AI systems in their actual business operations. This is where most of the complexity of the real world lies, and where a significant amount of value will be created in the next five years. Here is the irony that nobody talks about. Anthropic's Cowork - the product that allegedly killed SaaS - is itself a SaaS product, which is sold to organizations through subscriptions on the Internet. SaaS has always been a good delivery model. And SaaS, as a business strategy built on shallow moats and commodity workflows priced by seat, is already over. Conclusion: Everyone is watching this $285 billion loss and seeing destruction. But I see a transfer. That value has not disappeared. It is moving from companies that capture value by becoming intermediaries between humans and their tools to companies that capture value through execution, data, and infrastructure. The old script was to build a workflow UI, charge by seat, and increase revenue by increasing the number of employees for customers. The new script is to build something that has data, execution results, or protection systems. Pricing based on the value of delivery, rather than the buttocks on the seat. If you are the founder reading this article, the worst thing you can do is panic. The second worst thing is to continue building like in 2019. The best thing you can do is understand where value is moving - and stand where it is landing. The SaaS era is not over. The era of easy SaaS is over. To be honest, for anyone who is truly building something real, this is the best news in a decade.