The numbers are hard to ignore. More than 220 US unicorn startups (companies once valued at $1 billion or more) are now worth less than half their peak valuation. Nearly half of all 857 US unicorn startups haven’t raised a fresh funding round in the last three years. Startups that last raised capital in 2021 have seen their valuations fall 68% on average. Those that last raised in 2022 are down 52%.

This is not a correction. It is a culling.

CNBC reported this week that the $250 billion wave of investment that has flowed into AI leaders like OpenAI and Anthropic (ahead of what are expected to be history-making IPOs) has not just redirected venture capital. It has exposed the fragility of an entire generation of software businesses that were built for a world that no longer exists.

What Got These Companies Into Trouble

David Zhu, a former head of engineering at DoorDash, put it plainly: “All workflow-driven enterprise SaaS companies will be either disrupted or dead in the next decade.”

The specific structural problem is the per-seat pricing model. For years, the SaaS industry made its money by charging customers based on how many employees were using the product. More users meant more revenue. It was a clean, predictable model while human workers were the only way to run business workflows.

AI agents break that model completely. When a single AI agent can handle tasks that previously required a team of five, ten, or fifty humans, the per-seat logic falls apart. Companies are not going to keep paying for seats that nobody sits in.

The pre-ChatGPT generation of startups is caught in a bind. Their technology was designed around human workflows. Their revenue models assumed growing headcount. Their infrastructure costs were sized for a world where adding capability meant adding servers and engineers, not deploying a new AI model. Rebuilding all of that is expensive, time-consuming, and requires skills many of these teams do not have.

Some executives quoted in the CNBC reporting argued their companies need to “rebuild from scratch” to survive. Others said the transformation is too difficult and some will simply fade out slowly. Neither option is good news for employees or investors who backed these businesses at peak-2021 valuations.

The Software Stack Is Being Rebuilt Around AI

What is replacing these legacy tools is not always obvious yet, but the direction is clear. Enterprises are moving toward AI-native platforms where intelligence is built into the workflow, not layered on top of it. The new generation of enterprise software is being built around agents, not around dashboards and forms.

This is why the companies attracting capital right now are building for the agentic future. And it is why the companies that missed the transition are struggling to explain to investors why they should keep funding a shrinking market position.

The irony is that many of the pre-ChatGPT unicorns built genuinely useful products. They solved real problems. But the technology curve moved underneath them faster than their business models could adapt.

What This Means for Business

If you are running a business that still relies heavily on legacy SaaS tools from 2019-2022, this trend should be on your radar. Not because those tools will stop working tomorrow, but because the vendors behind them are under enormous pressure and may not invest in the capabilities you actually need going forward.

More directly: your competitors who have made the shift to AI-native workflows are now running operations with a fraction of the overhead. The gap compounds over time.

The businesses that are navigating this well share a few things in common. They have identified which workflows are genuinely automatable with current AI tools. They have built practical competence in working with AI systems, not just bought subscriptions. And they have started replacing the per-seat SaaS stack with purpose-built tools or agents that do the job at a fraction of the cost.

This is not a distant future scenario. It is happening in real businesses right now. The 220+ unicorns currently losing value are not hypothetical casualties. They are the evidence.

For business owners thinking about AI strategy, the question is no longer whether AI will change your software stack. The question is how quickly you want to get ahead of it versus how long you are willing to wait while your competitors move first.

If you want to understand what an AI-powered operation actually looks like in practice for your type of business, that is exactly what Enterprise DNA’s Omni Advisory service is designed to help you work through. No hype, no vague roadmaps. Just a practical look at what applies to your situation. Book a discovery call with Sam McKay to find out where the real leverage points are.