On-call electric shuttles for big campuses, adding autonomy one loop at a time
TL;DR Loopcart runs app-booked electric shuttles inside large private campuses such as universities, tech parks and big plants. It starts with trained drivers and adds low-speed driverless running only on fixed loops the campus owner controls. The campus pays a monthly service fee and riders ride free.
New Indian campuses are big and spread out. IIT Hyderabad sits on 576 acres at Kandi, outside the city, with about 4,475 students[1]. Hostels, labs, gates and canteens can be a long walk apart, and fixed-timetable buses either run empty or leave people waiting in the heat. Tech parks and big plants have the same shape, with a few busy peaks and long, thin stretches in between.
Most estate offices deal with it through a contractor's diesel bus or a handful of golf carts on informal routes. Nobody measures waits or trips, so nobody can tell whether the money is well spent.
I found no public data on campus trip volumes in India.
What it does, in regulatory terms: a private, on-campus passenger service. Autonomy runs only on roads the campus owner controls, at low speed, with a steward on board. Loopcart does not run driverless vehicles on public roads. Section 2B of the Motor Vehicles Act lets the Centre exempt vehicles for research, but only case by case[2].
Riders book a seat in an app or at a kiosk by the gate. Software groups requests into shared trips, so four or six small electric shuttles can replace one big bus on fixed loops. Every trip gets logged, which gives the estate office its first real picture of demand.
Once a loop has months of trip and obstacle data, Loopcart fits drive-by-wire and sensors to one or two vehicles and runs them slowly on that loop, steward on board, long before it removes the driver.
I've reframed the seed, which pitched driverless on-demand shuttles from day one. In India that would stall on the law and on safety sign-off. Selling a better shuttle service first pays the bills while the autonomy proves itself.
The rider is a student, employee or visitor and pays nothing. The payer is the campus owner - a university estate office, a tech-park facility manager, a plant or a hospital. Today they pay a bus contractor or buy carts outright.
I found no published rate for campus shuttle contracts in India, so the price has to be tested. My working assumption is a fee per vehicle per month that undercuts the current contract.
India has a public autonomy testbed now. TiHAN at IIT Hyderabad builds ground vehicles with its own drive-by-wire and rents testbed access to companies[3]. Abroad, one autonomous-shuttle maker has worked with airports since 2018 and runs projects on campus-like sites in Finland and the UK[4]. That gives an Indian operator partners to learn from and a safety story campus boards can follow.
Nobody publishes a figure for an Indian campus mobility market, and I didn't find one for campus shuttle spending either. The evidence is anecdotal - big campuses like IIT Hyderabad's 576 acres[1] and a growing set of tech parks.
That shuttle maker's airport partners[4] are a useful neighbour, because they show private sites buy autonomous shuttles before cities do.
TiHAN builds autonomous ground vehicles and offers testbed time, but it's a research hub and not an operator[3]. A foreign shuttle maker sells its shuttle for airports, campuses and hospitals, with no Indian presence on its site[4]. A UK firm supplies driverless software for industrial vehicles at ports and airports[5].
I found no Indian company running shuttles as a service on private campuses with an autonomy path.
| Difficulty | Hard (63/100) |
| Starting capital | ₹2.4–4.3 Cr |
| Time to launch | 12–18 months |
Running the attended service is ordinary operations work. The hard part is autonomy on Indian campus roads, where dogs, cyclists, cows and parked bikes all share the lane and none of them read the rulebook. You also need a safety case the campus board will sign, and drivers who stick around through the slow months.
Running a crewed shuttle on private campus roads needs no special licence. Driverless running is a different matter and needs legal sign-off first.
Building this outside India? The categories above usually have a local equivalent, so check with your own regulator.
Don't trust the autonomy stack to run without a steward on board until a safety case and legal opinion say it can. A model that looks fine on a quiet loop can still fail on a crowded one.
Want to work out the AI layer for Loopcart? Talk to Yash →
Here's what can go south -
Worth doing if you enjoy operations and have patience with institutions. If you only want to build the autonomy stack then I'd say it's a poor fit.
Prove the boring part first - one campus where logged trips rise and average waits fall below five minutes, paid for by the estate office.
Plan for roughly ₹2.4–4.3 Cr, covering electric shuttles, the booking software, drivers and stewards and the later autonomy retrofit.
Not for crewed shuttles on private roads. Driverless running needs a legal opinion on whether the roads are public, and possibly a section 2B exemption.
Yes, large campuses everywhere have the same problem. Autonomy rules vary a lot by country, so start crewed and check the local law before going driverless.
The free Loopcart validation kit has the line-by-line capital breakdown, the effort scoring, the vehicle-law and data-protection checklist, a 14-day validation sprint with pass/fail targets, a 90-day launch plan, 10 interview questions for estate managers and riders, and 5 assumptions to test. Download the kit →
Thinking about building Loopcart or something close to it? The make-or-break work is signing the first campus and writing a safety case its board will accept, and I can help you get both in shape. If you need a hand, you know where to find me. Talk to Yash →
For illustration only. Names are fictional and ideas hypothetical; sourced facts are real but may be out of date. Nothing here is legal, financial or professional advice, and Yash Khatri is not personally liable for any use of it. Read the full disclaimer.