Speed
Passport
- Rock
- Locks
- Depth
- 1 · Shovel
- Time to dig
- weeks — months
- Capital
- ○ · low
- Solo
- ✓ solo-reachable
- AI
- ↓ under attack
- Rent
- ✗ not for sale
Essence
Iteration tempo as a compounding advantage: while a competitor ships one version, you run five cycles of hypothesis, release and data. In the AI era "speed is the only moat" became a mantra; it is worth working out where that is true and where it is not.
How it is built
SpaceX — iteration against perfection
Build, blow up, learn, repeat, against the multi-year flawless programmes of cost-plus contractors: in the same period SpaceX runs an order of magnitude more learning cycles, and every failure is data it paid for. Speed here is not haste but the frequency of learning, built into an engineering culture and into vertical integration.
Shein — a cycle in days against an industry of seasons
From a spotted trend to a product on the page is about a week, with micro-batches against live demand instead of seasonal collections planned months ahead. Speed turns into structural consequences: almost no unsold stock, an assortment that breathes with demand — competitors on a long cycle cannot answer without rebuilding the whole chain.
Linear and Vercel — tempo as a public signal
Weekly visible releases become the marketing: an audience subscribes to the tempo itself, and investors and candidates read it as team quality. For small teams speed is also a hiring magnet — the fast attract people who hate the slow.
How it is bypassed
The tools of speed are commoditised
CI/CD, feature flags and now agents give everyone a fast lane, including yesterday's slow. The relative advantage of tempo is compressed by exactly the process that eats execution polish: when everyone is fast, speed is hygiene rather than a moat.
The giant's burst mode
Big companies can be fast when frightened: Google's code red after ChatGPT produced a catch-up line-up in months, and Meta cloned Snapchat's Stories across its whole audience in a quarter. A newcomer's speed beats a sleeping giant, not an awakened one — permanent sleep is not something to plan around.
Speed without accumulation is a window, not a wall
Tempo on its own leaves no sediment: if the lead is not spent growing structural assets — a data flywheel, an ecosystem, a community, a brand — the next fast player does to you what you did to the slow ones. The conclusion matches the one for execution polish: speed is the best wedge and the worst wall.
Verdict
Speed is a meta-moat: a multiplier for building every other moat and nothing on its own. It is built with a culture of frequent learning; it is bypassed by the commoditisation of the tools and by the burst mode of frightened giants.