To me d5 would be pretty much an automatic move. There might be a better idea in the position, but I don't see it.
The reasoning is as follows:
- Longterm black is more than ok, with Bg7, 0-0, Rd8, f5 …
- d4 is a weakness which also limits the Be3 and the Nf3.
- I don't have to worry about losing c5, because black is way too underdeveloped to grab a pawn like that.
Based on that I would be inclined to just play it without calculating anything. Then I would force myself to at least calculate one or two lines:
- 1.d5 cxd5 2.Qxd5 Nxc5 3.Bxc5 Qxc5 4.Re1+ and black loses her queen.
- 1.d5 cxd5 2.Qxd5 Rd8 3.Qc4 (black still can't take on c5 and I'm ready for c6, when blacks queenside falls apart.) 3…Qc7 4.Rac1 and I don't know how much this is, but black has to solve some problems (c6 is a threat. Qc6 Nd4 doesn't look good).
Basically, given that black needs at least two more tempi to get his king into safety, if you open the position with white nice things will happen.
Edit: Then I would make the move, write it down, get up from the board and look at the positions of my friends and then suddenly realise that black can also immediately take on c5, without taking on d5 first. That would give me a little jolt, but wouldn't worry me too much, because after all "black is way too underdeveloped to grab a pawn like that". But I would still hurry back to my board and calculate stuff like 1.d5 Nxc5 2.dxc6 bxc6 3.Bxc5 Qxc5 4.Re1 Be7 5.Rc1 Rd8 6.Qe2 Qb4 7.a3 Qb7 8.Rxc6 and relax.