The reason which was suggested is that the domain is a “kill switch” in case something goes wrong, but I now believe it to be a badly thought out anti-analysis.
In certain sandbox environments traffic is intercepted by replying to all URL lookups with an IP address belonging to the sandbox rather than the real IP address the URL points to, a side effect of this is if an unregistered domain is queried it will respond as it it were registered (which should never happen).
I believe they were trying to query an intentionally unregistered domain which would appear registered in certain sandbox environments, then once they see the domain responding, they know they’re in a sandbox the malware exits to prevent further analysis. This technique isn’t unprecedented and is actually used by the Necurs trojan (they will query 5 totally random domains and if they all return the same IP, it will exit); however, because WannaCrypt used a single hardcoded domain, my registartion of it caused all infections globally to believe they were inside a sandbox and exit…thus we initially unintentionally prevented the spread and and further ransoming of computers infected with this malware. Of course now that we are aware of this, we will continue to host the domain to prevent any further infections from this sample.