Uber didn't scale to 45 countries by being centrally clever. It scaled by being radically decentralised — and the founders I advise today who try to copy it usually copy the wrong half.
The model was simple in principle and brutal in execution. Each city operated like an independent startup. The city GM had real P&L ownership, real hiring authority, and real latitude to make calls that fit their market. London wasn't run like São Paulo. Lagos wasn't run like Singapore. The assumption baked into the design was that the person closest to the market understood it better than anyone at HQ — and they were right.
What decentralisation unlocked
Tempo. When you don't have to escalate every decision up three layers, you can respond to a competitor move on a Tuesday and have a counter live by Thursday. Most large organisations structurally cannot move that fast, regardless of intent.
But decentralisation has a cost that's easy to underestimate. Without strong shared infrastructure — consistent data, shared playbooks, tight feedback loops back to the centre — you get fragmentation. Cities re-solving the same problems independently. Institutional knowledge that stays local and never compounds. We learned that lesson in real time.
The version that actually works
Maximum local autonomy on what's genuinely local, maximum standardisation on everything else. Pricing strategy in a new city? Local. How you track driver supply? Global. What counts as a successful launch? Global. How you build relationships with local regulators? Local. Get those calls wrong and you've either built thirty-five disconnected startups or one slow, distant headquarters.
What AI changes
The cost of standardisation. A market launch playbook that took a country team six months now lands in two weeks. Localised pricing models, regulatory mapping, bespoke onboarding flows — most of it can be templated, generated, and tuned per market by a team a fraction of the size. The right operators won't use that to centralise decisions; they'll use it to push more decisions into local hands while keeping the centre coherent.
The cultural infrastructure mattered as much as the organisational one. The intensity at Uber in those years is hard to describe from the outside — high-pressure, often chaotic, deeply competitive. But underneath, there was a clarity of mission that kept people aligned across time zones and languages. Everyone knew what winning looked like. That shared definition of success was the invisible structure that made the decentralised model function.
The takeaway founders should hold onto: how you organise a company is a strategic decision, not an HR one. AI changes the economics of execution; it doesn't change the principle that structure compounds — for better or worse, faster than ever.