Key Takeaways

  • Public and late-stage private markets are discounting seat-based business models, pushing Figma down toward single-digit billions after Adobe's abandoned $20 billion acquisition.
  • Atlassian's warnings on slowing paid seat expansion show that enterprise headcount growth no longer guarantees automatic revenue expansion for legacy software.
  • Jason Lemkin points out that the proportion of knowledge work handled by humans is shrinking across both horizontal and vertical B2B software.
  • Rory O'Driscoll notes that agile platforms like Lovable, Replit, and Cloudflare move faster than incumbents because their architecture assumes software gets written by agents rather than human engineers.
  • Lemkin argues incumbents must sacrifice gross margin basis points to run compute-heavy autonomous workflows, or risk losing their entire customer interface.

The Math Behind Per-Seat Expansion Just Broke

For fifteen years, cloud software ran on an easy equation. A company hired twenty engineers or designers. That company bought twenty licenses. Next year, they hired forty more, and the software contract doubled without the sales team lifting a finger. Net revenue retention stayed above 120 percent because white-collar payrolls kept growing.

That compounding loop is stalling. Enterprise customers are freezing or trimming headcount while software agents soak up the workload. As Rory O'Driscoll observed, public and private markets are waking up to this shift: “The new trend is agents not engineers and that means the seat license thing is brought into question.” (play from 55:06) When Atlassian warned Wall Street about slowing user growth, it was not an isolated sales execution problem. It signaled that companies need fewer human seats to run Jira tickets.

Figma faces the same pressure. After regulators blocked Adobe's $20 billion acquisition, Figma saw its private market valuations drop toward single-digit billions. The risk is not that designers prefer a rival drawing tool. The risk is that teams will bypass manual UI drafting entirely.

Incumbents Must Burn Gross Margins to Survive

Legacy SaaS built its business on 80 percent gross margins. Running simple CRUD databases and rendering vector shapes in a browser costs almost nothing in compute. But running autonomous agents that generate complete production code and interactive user flows burns expensive GPU hours.

Jason Lemkin warns that incumbents are clinging to high margins while younger competitors eat their core workflows. Lemkin put it bluntly: “I truly believe if you look at all the most impacted companies, there are seat-based ones that the markets do not believe will benefit from agents and AI. And I believe Figma, unless it radically changes, it is going to be in the doldrums.” (play from 58:48)

His prescription requires self-cannibalization. Lemkin argues Figma must take on the cost of building autonomous end-to-end applications: “I would go all in on agentic design. We will build you agentically your app. Every page, every element of it will be as good as Notion or better using our AI and our agents, and it is going to cost us a lot of basis points.” (play from 1:00:34)

O'Driscoll agrees that defense is a losing strategy: “It really is march or die. You either get with the program or you're toast, and it's not even interesting to say it anymore.” (play from 51:33) Startups like Lovable and Replit are not selling tools to help humans type code five percent faster. They are selling the output directly. If incumbents refuse to absorb compute costs, autonomous agents will generate assets outside their walled gardens and render their editor seats irrelevant.

What to Do With This

Audit your pricing page this week and count how many line items charge per human user. If seat expansion generates more than half your expansion revenue, draft a usage-based or outcome-based pilot tier that charges for tasks completed rather than logins created. Ship that pricing test to your next five enterprise renewal conversations.