Why investing in growth-stage AI startups is getting riskier and more complicated

Making a bet on AI startups has never been so exciting – or more risky. Established operators such as OpenAI, Microsoft and Google quickly scale their possibilities to swallow many of the offers from smaller companies. At the same time, new startups reach the growth phase much faster than they have historically seen.

But defining “growth phase” in AI startups is not that cut and dried today.

Jill Chase, partner at Capitalg, said on stage WAN AI sessions That she sees more companies that are only a year old, but have already achieved tens of millions of annual income and more than $ 1 billion of appreciation. Although those companies may be defined as an adult because of their appreciation and generating income, they often miss much of the necessary safety, recruitment and executive infrastructure.

“On the one hand that is really exciting. It represents this brand new trend of extremely rapid growth, which is great,” said Chase. “On the other hand, it is a bit scary because I am going to pay a rating of $ x billion for this company that no existence was 12 months ago and things change so quickly.”

“Who knows who is somewhere in a garage, maybe somewhere in this audience, starting a company that will be much better in 12 months than this in which I invest, today is $ 50 million Arr,” Chase continued. “So it has made growth a bit confusing.”

To cut the noise, Chase said that it is important for investors to feel good about the category and “power of the founder to adjust very quickly and see around the corners.”

She noted that AI Coding Startup Cursor is a good example of a company that “on the exact correct usage facility of AI code generation jumped that was available and possibly given the technology at that time.”

Cursor, however, will have to work to keep his lead.

“By the end of this year there will be AI software,” said Chase. “In that scenario, what cursor has today, becomes a little less relevant. It is an exhibition of the cursor team to see that future and to think, okay, how do I start building my product so that when those models come out and are much more powerful, the product surface that represents and I can transfer it to Code generation?”