Yeah, I think plastic composite coasters will become a thing eventually too. But I guess it'd take a decade or so after it has become a common material for use in other structures. And then some middle eastern mall with lots of money and high ambitions, but strict structural load restrictions, will shell out for a lightweight composite coaster structure for a family coaster. A few more of those to prove the concept, then finally they will start appearing at seaside parks where salt spray is a concern. Steel and wood will keep the largest shares of the market, though, because those materials are fairly cheap and durable.
Other than that, I suppose we'll see more implementation of audiovisual effects on coasters as well. Projection mapped images in tunnels that matches the speed of the train, for instance, or a ride soundtrack that strategically speeds up and draws out to always synch up exactly with the ride's pace. VR will be better, eventually, but I guess some indoor coasters will try to create a VR experience without the use of headsets too. We're sort of there with dark rides already, but those are hella expensive (the Harry Potter rides, for instance).
Oh, and the notion of coaster elements as we know them from RCT and older coasters will be gone before long. Arrow famously only made loops of one size back in the days, because that was the one design they had done the complex calculations for. By re-using those elements to exact specifications, they could re-use the costly calculations too. But now that computing power has increased so much, the designers don't have to stick to "tried and true" track shapes. Track can be designed and manufactured by powerful and precise computer algorithms, resulting in a smooth ride experience with just the right forces throughout, where one element glides seamlessly into the next. Vekoma's newer designs seem to utilize this philosophy, where it's hard to tell what's an element and what is a particularly twisted section of track.
Perhaps even entire park experiences will be computerized. Imagine a system where everybody gets a Q-bot, and an assigned ride time for their preferred rides, drastically reducing time spent in queues? Some parks are almost there already, so let's go one step up: Responsive park maps on your smartphone which tells you about which rides have the shortest queues, or which food outlets have the shortest waiting times right now. Digital park maps (or "Park guides", I guess) might also be used to give certain guests discount coupons for snacks and such, or Q-passes for low-traffic attractions, in an attempt to spread crowds out across the park, for instance if they find too many people are crowding in one section (if the attraction has low traffic, you wouldn't strictly need a queue pass, but the "If you hurry, you can skip the queue!" message is an incentive to try the ride out, at least). Heck, if enough guests registered for the system and preference data was stored for multiple visits, the parks could anticipate in advance where and when queues will form, before the park gates even open for the day, solely based on who shows up.
I guess things will also happen on the maintenance side, but that won't affect guests as much. For instance, after the accident with the Smiler, it would almost surprise me if Gerstlauer isn't looking into a system where an onboard computer on each train signals the position of the train in real-time, and locks down everything if the wheels are suddenly moving backwards. With some creative hooking up to the queue time estimating system, coasters could also tell on their own when it's time to put on another train. And of course, on-board train computers would be able to tell when it's time to lubricate the bearings or change the wheels, and accurately report the number and positions of guests on board in case of an emergency.
