Something strange is happening in software. The tools we used to pay $20 a month for, the scheduling apps, the form builders, the copywriting assistants, are quietly being absorbed into the platforms we already use. Not acquired. Not sunset. Just rendered unnecessary by a well-written prompt. AI is collapsing entire SaaS categories into features. And if your product can be replaced by a system prompt and an API call, you might not have a product anymore.

The great compression

The pattern is unmistakable. Anthropic launched Claude Design in April 2026, and suddenly non-designers could generate interactive prototypes from a text description, no Figma expertise required. AI coding agents like Cursor, Claude Code, and GitHub Copilot have turned what used to be full IDE feature sets into conversational workflows. AI scheduling tools are eating into Calendly's territory. Basic analytics dashboards can be spun up with a single prompt. This isn't hypothetical disruption. Venture capital investment in traditional SaaS dropped 41% year-over-year as of early 2026, according to a TechCrunch industry report. Bessemer Venture Partners, long the bellwether of cloud investing, publicly removed three SaaS categories from their State of the Cloud index, calling them "structurally uninvestable." Gartner predicts that 35% of point-product SaaS tools will be replaced by AI agents by 2030. The collapse is accelerating, and it's hitting the middle tier hardest.

From insult to identity

"GPT wrapper" was an insult in 2024. Founders would bristle at the accusation. By 2026, the uncomfortable truth is that most SaaS products are wrappers around an AI capability, whether they admit it or not. The line between "AI-powered product" and "thin interface on top of a foundation model" has blurred beyond recognition. Consider what a modern LLM can do out of the box: generate marketing copy, summarize documents, answer customer support queries, build simple apps, analyze data, draft legal contracts. Each of these capabilities used to justify a standalone subscription. Now they're table stakes, bundled into ChatGPT, Claude, or Gemini for the cost of a single API call. The SaaS companies that built their moats on feature differentiation are discovering that features are the most commoditizable part of the stack. When Anthropic added a legal task plug-in to Claude in early 2026, it wiped roughly $285 billion in tech market value within 24 hours. That's how thin the ice has become.

Who survives this

Not everyone is equally exposed. An AlixPartners analysis scored software companies on AI disruption risk and found that only about 14% had strong moats across both data ownership and vertical industry depth. Roughly a quarter had weak defenses on both fronts, leaving them highly vulnerable. The survivors fall into a few categories. Workflow lock-in. Tools like Notion, Figma, and Salesforce have something that a prompt can't easily replicate: they're embedded in how teams actually work. Migrating away from Salesforce isn't about finding a better CRM. It's about unwinding years of custom integrations, workflows, and institutional knowledge. The switching cost is the moat. Proprietary data moats. Companies sitting on unique, hard-to-replicate datasets have natural defensibility. This is why data infrastructure, security, and observability companies are actually getting bigger because of AI, not smaller. They own something the models need but can't generate. Systems of record. As Bain's 2025 Technology Report noted, systems of record sit at the base of the enterprise stack as the source of truth. They store core business data, manage access, and enforce compliance rules. Their edge lies in unique data structures, long histories of activity, and built-in regulatory logic that would be costly for anyone to replicate. If you don't fall into one of these categories, you're in the compression zone.

The funding paradox

Here's the irony that nobody in venture capital wants to talk about. AI startups attracted over $200 billion in funding in 2025 alone, capturing nearly half of all global venture capital. OpenAI and Anthropic have accumulated a combined $242.6 billion in venture funding. The money pouring into AI is staggering. But AI itself is eating the startup opportunity space. The very technology that investors are funding is making it harder to build defensible software businesses. Every new capability that Claude or GPT gains is another SaaS category that gets compressed. Every improvement in AI coding agents means a solo developer can replicate your product in a weekend. Investors are simultaneously building the bulldozer and funding the buildings in its path.

The floor is higher, the ceiling is lower

For indie hackers and solo builders, this era is paradoxical. The floor has never been higher. A single developer with AI tools can ship in 30 days what used to take a team six months. The barriers to building a functional, polished product have essentially collapsed. But the ceiling is lower too. If you can build it in a weekend, so can everyone else. The indie hacker who ships a clever AI-powered tool on Monday might find three competitors by Friday, each built by someone who saw the same opportunity and had access to the same models. As one Indie Hackers commenter put it: building is table stakes now. AI made that easy. But sales and marketing are still hard, still an art, still the thing that decides if you survive. Distribution, not the product, is the differentiator.

The laundry detergent problem

This situation has a useful analogy in consumer packaged goods. Walk down the detergent aisle and every product is functionally identical. They all clean clothes. The chemistry is commoditized. So what determines which one you buy? Branding, shelf placement, habit, and trust. SaaS is heading for its own detergent aisle moment. When AI commoditizes the core functionality, when every product can do roughly the same thing because they're all powered by the same underlying models, the winners will be determined by distribution, brand, and switching costs. Not by features. This is why the smartest SaaS companies are investing in community, content, and ecosystem rather than shipping more features. They understand that in a world where features are free, the product is no longer the product. The relationship with the customer is.

What this actually means

Let's be precise about what's happening. SaaS isn't dead. That's hyperbolic, and every few years someone declares the death of something that's very much alive. What's actually happening is more nuanced and more interesting. The middle tier of SaaS is being compressed. The single-purpose tools that charge $15-50 per month for one narrow capability are the ones getting absorbed. The platforms that own workflows, data, and customer relationships are fine, possibly even stronger, because AI makes their platforms more valuable. Some SaaS categories are genuinely expanding because of AI. Data infrastructure companies are booming because AI needs data pipelines. Security and observability tools are more critical than ever because AI systems need monitoring. Vertical SaaS in regulated industries has natural protection because compliance isn't something you can prompt your way out of. The question every SaaS founder needs to ask is simple: if someone described your product's core functionality to an AI model, could it do roughly the same thing? If the answer is yes, your distribution better be a fortress. Because your product just became a feature.

References