The maintenance/support/licensing costs of a given gizmo is probably roughly the same regardless of the cost or size of the physical box*. The software is becoming the bottleneck.
No, what to do with the thing(s) is the bottleneck. As projects like OLPC have shown, simply airdropping in technology and hoping it'll sort itself out is a recipe for little more than millions wasted on, well, airdropping in technology (unfortunately Ivan Krstic's "Sic Transit Gloria Laptopi" has gone 404 or I'd post a link to that).
About fifteen years ago I watched (from the periphery) as the US Govt sank several billion dollars into technology that was essentially an unproven hypothesis dreamed up by geeks (the fact that in the ten years before that no-one had been able to make it work didn't seem to bother them). A few years later someone published a report on it titled "Nothing but Pilots", which pretty much summed up the outcome.
On the one hand if you sit back with everything you do and wait for use cases to appear you'll never make much progress, but then you need at least one or two real, practical applications before you roll out some new geek dream. It sounds like the Barcelona deputy mayor has reached the "Nothing but Pilots" stage of the cycle.