Capital as a moat
Passport
- Rock
- Assets
- Depth
- 4 · Mine
- Time to dig
- 10+ years
- Capital
- ◉ · extreme
- Solo
- ✗ no
- AI
- → eroding
- Rent
- ✗ not for sale
Essence
Access to money a competitor does not have: patient investors, cheap debt, insurance float, cash on the balance sheet. It lets you subsidise prices, absorb losses longer than anyone else and buy up scarce resources. Product taxonomies do not tag it, but it decides market outcomes regularly.
How it is built
Amazon — investor patience as an asset
Twenty years of near-zero profit while investment in logistics and AWS kept growing: Bezos negotiated the right to reinvest everything, and competitors under a quarterly-earnings demand could not answer symmetrically. The moat is not the money as such but the mandate to burn it.
Blitzscaling — Uber and the subsidy wars
Billions of venture money spent subsidising both sides of the market until competitors died: it worked in the United States against smaller players and failed in China, where Didi was funded just as well and Uber left by selling the business. A capital moat only works where the capital is asymmetric.
Float and balance sheet — Berkshire and Apple
Berkshire's insurance float is free leverage for investing; Apple's hundreds of billions in cash mean prepayments to suppliers, capacity bought out and compute deals nobody else can make — capital converted into a cornered resource. The balance sheet as a weapon: not spending more, but being able to do what others cannot afford.
How it is bypassed
Efficiency outlasts money
The bootstrapped and profitable survive the winter: Zoho has grown for decades without venture capital against far better funded competitors, and Mailchimp reached a roughly $12 billion sale without a single round. When capital got expensive in 2022, the burners died and the efficient took their customers. Time plays against a capital moat: subsidies end (MoviePass), prices have to rise, and then unit economics compete rather than balance sheets.
Do not play where subsidies decide
Capital buys market share; it does not buy trust, regulatory precedent, community or depth in a niche. An attacker without money picks a field where money does not convert into victory directly, and digs slow moats there while the giants burn billions on each other.
AI compresses the need for capital itself
A product that used to need fifty engineers is built by three people with agents: the capital threshold for entry is falling faster than ever, and "we have more money for development" is ceasing to be a moat in software. Capital remains a moat where it buys physical things: compute, factories, licences, distribution.
Verdict
A cyclical moat: omnipotent in an era of cheap money and scarce resources, empty against the efficient and the patient. AI devalues it further in pure software.