The question should be, what would be the cost of transaction fees once we reach milestone.
Based on previous calculations, transaction fees (at that time) would be useless for dust/small transactions.
It largely depends on how scaling is resolved over the long term, along with the price. If Bitcoin's main chain achieves a greater capacity, the cost can be spread over a greater number of participants and fees shouldn't rise to the point where most users are unable to afford it.
If scaling is achieved predominantly off-chain, or through payment channels or sidechains, then some amount of whatever fees are being paid on those transactions might end up going somewhere else and not to the miners. In which case, the fee for securing transactions on the main chain would have to rise. This could potentially be prohibitive for those who aren't hodling large sums.