Stratfor's latest podcast (or here) addressed it in detail, but at root level it's very simple: the markets were pricing in (due to closeness of polling and other polling-related factors) the worst-case scenario of the two winners being right-wing Euro-skeptic Marine Le Pen and left-wing Euro-skeptic Socialist/part-time-Communist Jean-Luc Mélenchon.
In such a worst case scenario (from market's point of view) either of the two winners would be bad; so runoff round would be a choice between bad and worse.
The outcome of the first round was, instead, the much better scenario of market-friendly centrist Macron running (and, as per 538's analysis, most likely winning) against Le Pen; leading to Macron presidency. The markets like that it's not one of the two bad options in the first place, and Macron's general market-friendly policies and background don't hurt.
Aside from the above practical consideration, markets probably also reacted in part to the "well the Euroskeptic populism is now less on the rise than everyone assumed after Brexit and Trump".