Venture capital, the money that backs young companies, is flowing to fewer and fewer names. On a recent episode of the 20VC podcast, hosted by Harry Stebbings with Rory O'Driscoll and SaaStr's Jason Lemkin, one example stood out: a company with $20M in yearly revenue, growing 600-700% and profitable, could not get investors to take meetings.
Why good firms get ignored
The hosts pointed to math. A fund of $5B or more needs huge stakes to matter. Lemkin said that if a fund raises that much every 18 months, a $2B outcome does not work for it. Stebbings put it simply: unless a company can plausibly reach a trillion-dollar value, the top firms are not interested.
O'Driscoll said money is still going to odd places, such as data centers in space and quantum computing. What struggles is vertical software, meaning programs built for one industry, because of a fear that AI could make the whole category go away.
What founders were told
- Lemkin's advice: with top-decile numbers outside hot areas, expect to meet 100 investors.
- Someone will likely say yes, possibly at a high valuation.
- Strong numbers used to mean an automatic offer. That is no longer true.
The risk in the AI trade
CoreWeave, a company that rents out AI computing power, fell 15.5% on October 14 to close at $65.78. That was its biggest one-day drop since February. Some of its bonds traded at yields above 15%. O'Driscoll said such firms carry heavy spending and lots of debt, so they have little room to cut back when demand wobbles.
Lemkin called CoreWeave and Oracle second-tier next to Nvidia, and said such assets react harder to shocks.
Anthropic's big quarter
Anthropic told prospective investors its third-quarter revenue was $16.8B, up 45% from $11.6B in the second quarter and more than triple the first quarter's $4.7B. Some attendees discussed a valuation near $1.5T, below the roughly $2T investors had expected for the November IPO. O'Driscoll called the slowing growth normal from such a high level. Lemkin warned the stock could be choppy, since developers are not loyal and market share has swung a lot.
What happens next
O'Driscoll noted that history shows 10-20 decacorns, companies worth $10B or more, per decade in US tech, while the market acts as if there will be 90 this year. He is betting on the history. If demand growth slows, he said, price cuts among AI firms could become the main weapon.
The takeawayOutside hot AI areas, expect to pitch many more investors, and treat debt-heavy AI stocks like CoreWeave as high risk.