The usual definition of "classic" is something that is the best of its era. But I don't think that really applies to instruments. There are intangible factors.
Intangible #1, Historical Significance: The DX7 is considered a "classic," but I don't think it's one of the best synths of its era, or anywhere near it. But it has some historical significance as a groundbreaking and genre-defining digital instrument, so it's considered a classic. Contrast this with the Korg Polysix, which did almost nothing new, but is still a classic.
So you might have some instruments like that. The Voyager comes to mind, as the kickoff of the new Moog Music. Perhaps even the Little Phatty, as Bob Moog's final instrument (even though he's not the one who finished it). The Evolver, certainly, as the kickoff of the new Sequential/DSI. Historical perspective is really hard to pin down.
Intangible #2, Suppy-and-Demand: The Korg Polysix is considered a "classic," but I don't think it's one of the best synths of its era, or anywhere near it. But it was built poorly. The wood parts were a crumbly medium-density fiberboard, and the conductive rubber contacts failed pretty reliably. So if you manage to find a Polysix in somewhat good shape, it's going to be worth a lot. Contrast this with the DX7, which is built like a tank, and isn't particularly valuable due to its vast numbers.
Paul Dither already mentioned the Volca series as future classics, and I agree with that. They're dazzling as musical instruments, but they've got thin scratchable plastic bodies, battery covers that are going to get lost and broken, flimsy knobs, and plastic jacks. They're not exactly built for the rigors of the road. So in ten years, the ones that survive are going to be sought-after and expensive.
This brings me to a crazy prediction that the Desktop Evolver might wind up being an expensive classic. It basically hits all three points: A groundbreaking instrument in its day, with historical significance, and with build quality that, while not atrocious, leads me to expect long-term attrition.