What an operational moat looks like in autonomous mobility
In software, moats compress. The defensibility of any pure-software AV company three years from now is mostly determined by who has the cheapest GPUs and the best PhDs. In operations, moats compound. The defensibility of an operational layer ten years from now is mostly determined by who started building it five years earlier. Autzu is built around five overlapping operational moats — each individually replicable, none of them simultaneously.
Moat 1 — Physical real estate
Every autonomous-ready hub requires utility-scale electrical service, zoning permission for 24/7 fleet operations, traffic access to a service polygon, and 30,000–80,000 sq ft of land within commute distance of dense urban demand. In every market Autzu enters, those sites are individually expensive to acquire and impossible to acquire fast. Once you have a site with a 20-year lease, a competitor catching up means finding equivalent land in the same metro — typically a 2–3 year endeavour, often longer.
This is the moat people underestimate because it looks unglamorous. It is the moat that ultimately decides the market structure.
Moat 2 — Utility relationships
A hub of 250 robotaxis requires 3–7 MW of utility service. In every market, that requires a substation upgrade or transformer placement, which the utility schedules 12–24 months in advance. The operators that began coordinating with PG&E, Hydro One, SRP, ERCOT, Toronto Hydro, and the equivalents three years ago are now positioned to switch on. The operators that begin coordinating today will be live in 2028.
Utility coordination is non-replicable on a budget. It is replicable only on a calendar.
Moat 3 — Software learning
Every charger handoff, every cleaning cycle, every calibration session, every dispatch decision feeds back into the platform. After a million hub-hours of operations, the software knows things that a competitor cannot synthesise from first principles: where energy bottlenecks form, which vehicles drift on which sensors, which dispatch packings are unstable in which weather. This is the AWS-style learning curve — the more it runs, the cheaper it gets.
Autzu already has hub operational data spanning multiple years across multiple markets. A competitor starting today is years behind on data, which translates directly into cost per mile.
Moat 4 — Partner integrations
Real fleet operations require integrations with vehicle OEM telematics, AV stack vendors, payment processors, insurance underwriters, rideshare platforms, energy retailers, utility billing systems, regulatory reporting, and municipal incident response. Each integration takes 3–12 months and is contractually slow on both sides. The cumulative integration surface is, in practice, the most underappreciated barrier to entry in the entire AV stack.
Autzu's integration surface today includes Uber, Tesla, and several enterprise fleet platforms. The cost to replicate this from scratch is measured in tens of millions of dollars and 18+ months.
Moat 5 — Regulatory and permitting expertise
Operating fleet infrastructure across multiple jurisdictions requires fluency in the California PUC, Toronto PTC, Arizona TNC, Texas TNC, NYC TLC, and the dozens of municipal noise/zoning/traffic regimes that apply on top. Every market is its own permitting puzzle. The operator that has cleared three markets has a permanent head-start in clearing the fourth, because the documentation, the relationships, and the workflow already exist.
The compounding property
No single one of these moats is impossible to replicate. The compounding property is what matters: a competitor must replicate all five, in the same markets, on overlapping timelines. The cumulative time-to-replicate, even with unlimited capital, is 4–7 years. That is the durability that capital is buying when it invests at the infrastructure layer rather than the application layer.
References & further reading
- Hamilton Helmer — 7 Powers: The Foundations of Business Strategyamazon.com
- CB Insights — The state of autonomous vehicles 2025cbinsights.com
- Stratechery — Aggregation theory and infrastructure moatsstratechery.com
- a16z — The market map for transportation infrastructurea16z.com
