I think Zoom has become critical infrastructure, but not from a technological perspective. As @nicoburns points out in another comment, there are many companies

I think Zoom has become critical infrastructure, but not from a technological perspective. As @nicoburns points out in another comment, there are many companies that can do the job (MSFT, Google, Cisco, etc. have perfectly comparable alternatives here), and if you're willing to look outside of the biggest companies explicitly targeting business meetings, you have other companies with working technical infrastructure targeted at a different market (i.e. FB, Appl) and a batch-ful of YC startups trying to provide the same functionality to the same market. What I think the article is pointing out is that when Zoom goes down, schools can't teach. And that's true. I think this is indicative of a usability / coordination / culture problem of missing fallbacks. For example, when Zoom goes down a teacher (at least to my knowledge) can't/won't quickly email out a "Hey class, Zoom is down so here is the Google meet link for today". However, this is true of many technologies (if gmail is down I'm not getting email, if slack is down I'm not reading messages, and if Github is down I'm not pushing code). I think the larger problem is HOW MANY institutions it's a critical point of failure for. If Zoom goes down, our courts largely pause, our education largely pauses, our economy largely pauses, etc. The same is not true of Github or Gmail. In this regard it's closer to a powergrid failure. This should go away as verticalized solutions and fallbacks appear (leading to broader diversity, where not all courts use the same platform), but I think the key things to think about until then are: 1. Is there anything Zoom could do to create stronger network effects (https://ayushsood.com/zooms-missing-platform/ https://ayushsood.com/zooms-missing-platform/) preventing new entrants from entering the market? 2. If they did those things, how could we respond to ensure Zoom's incumbent position didn't allow them to prevent new entrants? An old-internet answer would be to centralize the key nodes, effectively utilitize them, and then allow anyone to make a platform that connects to them (i.e. AT&T doesn't own telephony, DNS' don't own the internet, Gmail doesn't own email). I'm not sure how likely this is given (1) we're moving towards centralization generally (2) the importance of performance in video streaming. Any abstraction creates overhead, and that might be a critical failure to open standards in the video streaming space.