Name: Starship delivery robot
Brand: Starship Technologies

Who's exposed

Jobs in the threat radius

food delivery courier (gig platform rider)
campus dining delivery staff
grocery last-mile delivery driver

Deployment status

The largest sidewalk-delivery operation in the world. In April 2026 Starship passed 10 million cumulative deliveries, with 3,000+ robots running daily across 300+ locations in 8 countries and 65+ US university campuses. A $50M Series C (October 2025) funds expansion toward 12,000+ robots by 2027. In November 2025 Starship signed a global partnership with Uber Eats: live in Leeds, UK from December 2025, multiple European countries through 2026, US in 2027.

When this hits the labor market

On campuses and in planned suburbs, substitution is present tense. Starship says a robot drop already costs $3-4 less than a rider-based delivery, with a long-term target near $1. Over 1-3 years, expect short-radius orders in US and European campus towns to keep shifting from couriers to robots — and the Uber Eats integration puts that shift inside the very platform couriers work on. The 3-5 year question is city sidewalks and permitting. The ceiling check: robots and drones combined still carry under 1% of deliveries today, with Barclays projecting ~10% by 2035. The direction is settled; regulation sets the speed.

Ten million deliveries is not a pilot

In April 2026 Starship posted a number nobody else in sidewalk delivery can match: 10 million cumulative deliveries. The fleet behind it is 3,000+ robots across 8 countries, 300+ locations, and more than 65 US university campuses. Forbes put the cohort comparison plainly last October: Starship’s volume is roughly five times all of its US competitors combined.

The machine itself is deliberately boring. Six wheels, three grocery bags of capacity, walking pace, Level 4 autonomy on sidewalks with more than 99% of driving untouched by humans. It runs 18 hours on a charge and parks itself on a wireless pad overnight. Nothing about it is humanoid, and that is the point — it only has to beat a courier on cost and reliability, not on shape.

The unit math already works

Starship’s CEO has said the quiet part: no tips, no gas, no wages. The company reports a robot delivery already runs $3-4 cheaper than a rider delivery and is profitable per-drop in some markets, with a long-term cost target around $1. Investors funded the scale-up in October 2025 with a $50M Series C aimed at growing the fleet from 2,700 to more than 12,000 robots by 2027.

The contract that should worry couriers landed in November 2025. Uber Eats integrated Starship robots into its own app — Leeds first, Europe through 2026, the US in 2027. The same order pool gig riders compete in now includes a row of six-wheeled bidders that don’t take a cut.

Why we track it

Most robots in this catalog are still proving they can do the work. Starship is proving how much of the market the work can take. Its chosen terrain is shrewd: campuses and tidy suburbs, where right-of-way is clear, orders are dense, and ticket sizes are small — precisely the orders that pay couriers worst. The robot absorbs the lowest-margin deliveries first, and the platform reprices around it. That is the standard substitution path in gig economics, running on schedule.

Urban peak-hour orders stay human for now. But if your income depends on campus-adjacent or suburban delivery runs, this is not an experiment somewhere else. Barclays estimates autonomous delivery could unlock $16B a year in platform profitability. Today, most of that line item is called courier pay.