The problem with this, which I learned the hard way with my first startup, is that roaming costs are not high because of some horrible inefficiency in the telco's charging them, but because they all benefit from the status quo.

My first company was an ISP, and we thought the then current dial up prices (1995) were insane, and quickly worked out we could offer packages at 25% of the largest competitor and make a profit with only 1500 or so users. Problem was, the largest competitors could also make a profit at 25% of their then-current prices, and had large established user-bases. We were inexperienced enough in business to not worry too much about why their prices were to high - we assumed without thinking that their prices reflected their cost base. Instead it reflected their belief about the market at a time - they'd rather sell at that price with a ridiculous margin, than cut the price and grow the total size of the market.

We went in cocky about a big competitive advantage, and it took about 3 days after launch before the incumbent dropped prices to match ours, and while we kept growing, it obviously made things a lot harder.

This company might find the same in the mobile data market if they get any traction at all: Many of their competitors can drop the roaming charges massive pretty much over-night, have large warchests, and have roaming agreements in place that can quickly be re-negotiated if they and their partners start seeing this startup as a threat to their revenues. The massive benefit they can provide now, and the seeming headstart they have on their competitors can evaporate pretty much overnight.