The New Moats
Jerry Chen · 2016
"Systems of intelligence built on proprietary data and AI/ML form the durable competitive moats of the next era."
A 'moat' is whatever protects a business from being copied or out-competed. Older moats were things like distribution deals, brand, or economies of scale. Chen's argument: in the era of cloud and AI/ML, the new durable moat is a 'system of intelligence' — a product that continuously collects proprietary data from how it's used, feeds that data into machine learning to get smarter, and becomes better and more entrenched the more it's used, in a way a new entrant can't just copy by writing similar code, because they lack the accumulated data and usage history.
The mechanism is a data-driven flywheel: usage generates unique data → that data trains models that make the product smarter/more personalized → a better product attracts and retains more usage → more usage generates more (and better) data. A competitor can copy the interface or even the underlying algorithm, but they can't copy years of accumulated proprietary usage data — so the moat isn't the code, it's the compounding data advantage earned through actual customer usage over time. This is different from older 'systems of record' (databases that just store information) because the value here comes specifically from the intelligence layer built on top of that data, not the storage itself.
Why does Jerry Chen argue that a 'system of intelligence' is a stronger moat than older forms of competitive advantage?
Read more about the topic
The explanation above is written with AI assistance. These are the originals — go to them to check it.
- The New Moats (original essay)Greylock — Jerry Chen
- The New Moats — Slides Edition Remix!Greylock Perspectives
- Venture Confidential, feat. Greylock's Jerry ChenHeavybit podcast
Do Things That Don't Scale
"Early startups should manually recruit users and do unscalable work; growth comes from effort, not automation."