Two games, both on this side of the Atlantic (and Pacific, for that matter). Vangers: One for the Road, and Remember Tomorrow.
Vangers was a very unorthodox (you could even say a bit psychotic) take on the combat-trade-sim genre, with a universe so full of alien concepts (and perfectly normal concepts under ridiculous names) that it made immersion a tad too difficult. Despite that, it was, in its own way, a brilliant game with lots of innovative concepts (like realtime terrain deformation in a sandbox vehicle-based game, way ahead of its time and still not quite replicated today), and really deserves a more coherently structured and better planned sequel.
Remember Tomorrow was a 4X RTS, with smart pause. The global time would pause whenever there was combat, and whenever you went to design a ship. And time speed could be varied, too.
The game was riddled with problems, like one race having no Dreadnouts despite having them in the research tree, and the economy screen being somewhat incomprehensible when there were too many planets conquered. Despite that, it shined through with two primary things. Ship design and technology research. The ship design was quite unlike anything I've seen pretty much anywhere else, before or since. You are given an outline of the ship's hull, within which there is a grid. Into this grid you must squeeze rectangular profiles of the various components available to you, like some deranged Tetris variant. Sometimes, the biggest benefits of a new piece of technology isn't its increased power or reduced cost, but the fact that it has a slightly slimmer or a differently oriented shape, thus allowing to fit it into nooks and crannys of the ship hull where you couldn't put them previously. The Earthling Frigate is a big example, where prior to researching Neutron Generators you had to occupy the dorsal external main turret mount with a 6x6 Fusion generator, and afterwards you could power the whole thing with an array of three 6x4 Neutron Generators stuck into the ship's nose section, with plenty of space left in the aft for fuel tanks and engines, and plenty of power to put four fast-tracking blaster turrets into the main dorsal mount.
The Research system was less innovational, and could easily be improved, but I still think it's one of the best research systems made for 4X games so far. Basically, you have one huge slider to distribute available scientists between General and Applied research, four smaller sliders to determine how many scientists work on general Physics, Mathematics, Biology, and Chemistry areas, and a big list of applied science areas that you can allow or disallow to focus research into a given direction. There were no "projects", all discoveries were a side-effect of amassing sufficient knowledge in specific areas. You could see the requirements for some things if you had all required applied fields open, but the only way you could focus on specific goals was by disallowing research in unneeded branches, pooling all research power into the required areas. I thought that was a great system, which could be easily improved by adding variability to the requirement values, and introducing a few more general areas (Psychology and Ethics, for example) as well as some more complex requirement schemes. And improving the interface, the one they used was very broken.
I hope that one day there's a game that takes at least these two key aspects of Remember Tomorrow, and improves on them, finally making a space 4X game that I can really enjoy.