Uber offers a lot more than just locating a driver. It’s a platform that generates a recommended route for the driver, shows the progress of the route to the customer in realtime so they know the driver is going the right direction, it calculates fares (e.g. distance, surge pricing, etc), and has other features like “share my ride details with my spouse” for example…for security purposes.

That’s just what i could think of off the top of my head and i’m not a product manager at uber. Things are NEVER as simple as they initially seem.

Other thoughts:

- at uber’s scale, payment processing alone likely takes up quite a bit of compute server side. You could outsource this to a 3rd party, but it’s still an expense as you scale.

- how would a decentralized system work if something goes wrong during the ride? who does the customer call? Drunk driver, attempted rape, etc. list of things that can go wrong is long.

- your platform/app would need to be smart enough to know where all the “proper” ride share pickup spots are at every airport. Not trivial!

- how would you offer a feature to “reserve a ride” like uber does, if you only build an app that can find “nearby drivers”?

food for thought. :) All im pointing out is this kind of product needs to be a platform and is likely not cheap to run. It’s not 90% client side. Unless you literally just want to reinvent the taxi concept, in which case maybe the 90% client side approach is feasible.