It's free for the same reason everything is free these days. VC funds anything that will attract a lot of users to mine data from so they can sell the data. Discord didn't do anything that was groundbreaking or even solve a problem that had no solution; they just came along during a time when investors are willing to fund a company operating at a loss for a decade until FANG buys them.
Discord's a pretty good product, and they've got the engineers and money to get better, but the only reason they won is because of timing. Same for Slack; there were identical products to Slack that tried for decades to gain traction, but they weren't free, because that business model didn't exist at the time.
The ux of Slack is essentially screen+irc implemented in JS with emotes. It enabled technical and non-technical people to use the same tool. The key to success is not technical, it's that they tailored the product to a specific group that would then lock itself in.
I didn't understand Discord's success, but comments here point that gamers couldn't find free group-voice apps at a critical time. Here again, they tailored the product to a group that would then voluntarily lock itself in.
Later, they sell the companies with valuations based on the captured user bases.
I wasn't referring to things like IRC. When Slack was initially released, it was no different from Campfire and a whole string of other web-based chat systems that came and went going all the way back to the dawn of AJAX in the late 90s. Slack's improved a lot since then, with app integrations and other features, but fundamentally it wasn't any different than its predecessors. It's easy to think that Slack did something groundbreaking, or figured out the magic solution to the problem that sank its predecessors, but just like Discord, the reason Slack won is because it came along at a time when companies can raise tens or hundreds of millions of dollars to float them for years while offering a free product. Then they can upsell later, and/or commoditize their users' personal information. Those business models weren't as easy to come by in the past, so a lot of products failed. None of this is to bash Slack; it's an adequate product for what it does.
Another big thing that the current crop of winners has going for it is that cloud hosting allows applications to launch literally for free and scale quite a bit without paying much of anything in infrastructure costs. That also wasn't an option 10-20 years ago.
Discord's a pretty good product, and they've got the engineers and money to get better, but the only reason they won is because of timing. Same for Slack; there were identical products to Slack that tried for decades to gain traction, but they weren't free, because that business model didn't exist at the time.