Somebody has to tell the room the launch is delayed. Whoever does it will be remembered slightly differently than everyone who already knew and said nothing.
What the research found
Leslie John, Hayley Blunden and Heidi Liu ran eleven experiments on exactly this. People who simply delivered bad news were liked less than people who delivered good news, even when the messenger had nothing to do with causing it. People who were merely present for the same news were not penalised. The listener, it seems, reads a motive into whoever says it.
These were mostly online studies with American adults, not offices, so take the size of it loosely. The direction is the useful part. It also explains the urge to put off saying it. That urge is a response to a real social cost.
Why the cost is real but manageable
This part comes from practice, not from a study. The cost is why how the news is delivered matters as much as the news itself. Stated plainly, with the relevant context attached, and paired with what happens next, bad news is received as information. Delivered vaguely, defensively, or with visible discomfort, the same news gets received as a verdict on the person saying it.
Where to start
Before delivering something difficult, separate the fact from your own reaction to delivering it. State what happened, state why, and state what comes next, in that order, without apologizing for being the one saying it.
The discomfort belongs to the news. It does not need to attach itself to you as well.
Reading how much context a specific piece of bad news actually needs, versus how much would just delay the point, is a judgment only you can make for the room you are in. What this gives you is the order that tends to keep the news separate from you as its messenger.
There is a specific way to deliver bad news that leaves the room focused on what happens next rather than on how you said it, and it is worth having ready before the next time you are the one who has to say it.