there seems to be a “rule of 40” for AI businesses
if you are growing >3X yoy investors seemingly accept negative to low GM and ignore questions about durability
or, you can grow “fast” but not insanely, and you need a much better moat (models, enterprise)
weak on both = tough
Positive users find Sarah Guo's Rule of 40 for AI growth and moats easy to accept since it feels about right, while negative users criticize it for depending on unrealistic 3x growth that masks weak unit economics and margin issues.
implied here is a belief in the value of scale (as one rough swag, reach of $500M+ ARR to enterprises or consumers)
there seems to be a “rule of 40” for AI businesses
if you are growing >3X yoy investors seemingly accept negative to low GM and ignore questions about durability
or, you can grow “fast” but not insanely, and you need a much better moat (models, enterprise)
weak on both = tough
Interesting how this post illustrates founder vs VC thinking.
VC: thinking about how other VCs think about businesses
Founder: thinking about the business and how to improve it
I think Sarah also goes into founder-mode quite a bit, but I nonetheless find the founder-VC distinction revealing.
@saranormous buried beat is rule of 40 prices growth or moat as a binary and skips what decides which side you hold. cost-to-serve under inference load is the real switch. capital allocators start scoring on unit economics by quarter end.
@saranormous Whats the best way to keep an eye on these new up and coming AI businesses? Is there an easy way to see what ideas are in play?
@saranormous what stage of businesses do you mean here?
everything i’ve seen pre-series c is raising based on month-on-month growth…
@saranormous Same dynamic in real estate. Flip 30 houses a year at thin margins and lenders throw money at you. Do 5 a year, your numbers better be bulletproof or nobody picks up the phone.
@saranormous Durability feels like the missing half of the agent conversation. Fast growth gets attention, but if an agent keeps identity, memory, permissions, and work history for years, the moat looks less like one workflow and more like a network. That's the bet behind SilicoVille.
@saranormous Where does the “40” come from?
@saranormous Feels about right, what do you think breaks first, growth or moat?
@saranormous There is no "insanely" but chemical constraints unlocked not by the exogenous
@saranormous 3X growth covers a lot of sins. The question nobody asks at 3X is what the unit economics look like at 1.5X.
@saranormous Feels right, but I’d argue the market is really rewarding speed plus proof now. Fast growth can cover a lot, but it doesn’t replace durability, especially when inference and infra costs are still hitting margins hard.
@mjzellinger @saranormous Depends on the VC. CVCs think partly (if not entirely) about strategic fit with internal business units, and many institutional VCs are willing to be contrarian and have their own views. But those tend to be specialist firms, often with teams having relevant domain background.
@saranormous honestly curious where the line is between "fast" and "fast but not insanely"
Health AI rewrites the rule. Durability comes from the regulatory and clinical moat, not just scale. We can stay sub-$100M ARR for 3 years building 6-track FDA + CMS reimbursement architecture, then go vertical when no competitor can clone it. Scale is downstream of trust in this market. Building it now at HabitCure — 6 patents on the detection layer, $3M seed closing.
@saranormous Good morning sarah 💙
@saranormous Scale without margin discipline tends to resolve one way. SaaS ran this exact playbook. Revenue isn't profit.
@saranormous ngl the 3x yoy is doing a lot of heavy lifting there
everyone screams AI but 99% of those companies arent hitting that
@saranormous Easy!