$20M ARR, Profitable, Growing 6-7x — And No VC Will Meet

By Grace Kim

A software company with $20 million in ARR, growing 6-7x and profitable cannot get a meeting with a top-tier VC firm. Jason Lemkin says that story, surfaced on this week's 20VC x SaaStr recap with Harry Stebbings and Rory O'Driscoll of Scale Venture Partners, defines the current market better than any dataset.

"That world is gone if you're outside the 3-4 categories everyone wants," Lemkin wrote in his weekly SaaStr column, Jason's Takes. His advice for founders outside AI infrastructure: plan the round like it's 2009. Build a list of 100 investors, not 15. Budget 2-3 months. Someone will still fund the deal — just not Benchmark, Index or a16z.

The week's other numbers set the stage: Anthropic booked a $16.8 billion quarter, CoreWeave fell roughly 25% in a week, Modal raised at a $15.75 billion valuation, and Berkshire Hathaway put $330 million behind Bending Spoons, the serial acquirer that recently bought Miro and Airtable.

What does fund math have to do with your rejection?

A $5 billion fund raising every 18 months needs billion-dollar positions — billion-dollar gains on its stake, not just billion-dollar exits. In the investment committee memo, a $1B gain is the B case; the A case is $25 billion or more.

So a founder pitching a $2 billion outcome to a mega-fund isn't being rejected on quality. "You're being rejected on arithmetic," Lemkin wrote. His rule: divide the fund size by 20 before every pitch, and ask whether your realistic outcome produces a position that size.

Is one-and-done fundraising still viable in the AI era?

Lemkin used to champion the single-round model — raise a couple million, never raise again. Zapier remains the canonical case: one real venture round, about $300 million in revenue, probably never an IPO.

But he now calls it a trap. Once VCs own the cap table, dividending out profits gets hard. If an investor owns 30% and you buy them out at 6x cash flow, you've tied up two full years of cash. The alternative is 37signals, where founders split roughly $200 million a year because they never took venture money.

"None-and-done is now realistic for a lot more founders than it was three years ago," Lemkin wrote, citing AI's ability to shrink the team needed to reach $10 million ARR.

How should founders pick seed investors?

Lemkin highlighted a portfolio company that raised about a $3 million seed, plans to never raise again, and is growing 15-18% a month, fully agentic. Still, nobody wants to meet. That means seed investors must be able to carry it with reserves — and many can't.

His two screening questions for every seed fund: What are your reserves per company? And what did you do the last time a portfolio company was growing well but couldn't raise?

What can the Anthropic and Replit stories teach about pricing?

Anthropic was expected to IPO at $2 trillion; the talk dropped to $1.5 trillion. Chamath Palihapitiya's take, which Lemkin endorsed: raise whatever you need at $1.5 trillion and move on. "It's just a valuation. Win the day." The same logic applies down-market — pick the valuation you can comfortably grow into in 18 months, not the one that tops TechCrunch.

On models, Replit flipped its default from Anthropic's Opus and Fable to OpenAI because OpenAI was 30-40% cheaper for roughly the same output. Lemkin's team uses Replit 14-15 hours a day; the press called it a breakup, he didn't. Notably, Replit, Lovable, Base44 and v0 — each around $1 billion in revenue with cost-sensitive customers — none moved users to open-weight models like Kimi.

That routing math matters: a 30-40% COGS swing on the same quality separates a 50% from a 70% gross-margin business. Lemkin also warns AI revenue is less durable than classic B2B ARR — Anthropic's revenue share could fall 10% in a year while the business stays great in absolute terms — and says to model 20-30% of usage revenue as at risk in any 12-month window.

Where would Lemkin put money today?

Build-to-prod. Generation is "solved-enough," he wrote, but reviewing, testing and hardening AI-generated code still takes an order of magnitude more time than generating it. Claude Code Review's launch proves the category rather than ending it. The adjacent warning: don't build a company that exists because someone else needs it for now — CoreWeave dropped 25% in a week on exactly that risk.

His closing frame: the venture market has split into mega-funds that need $25B outcomes and everything else. Founders in the second group, Lemkin argues, now have more options than ever — including skipping venture entirely.