A great question. I'd love to see a long-term version of this, too.
@das_gug already laid out the main historical comparisons. No previous blackout has ever affected more than 50M, and only one has affected that many, although as @404_NOT_FOUND points out, if the 2003 event were to happen again today, it would affect >60M because of population growth. More than half of US households had electricity as of 1925, and we've seen this one large event in the century since then. So, purely based on history, I'd put the likelihood at ~1%.
Such a blackout would require power to be out over a very large area. There are 2 main grids in the US and Canada: the Western Interconnection and the Eastern Interconnection. In addition, there are the smaller Texas, Quebec and Alaska Interconnections. The Western Interconnection serves a population of >80M: https://www.wecc.org/epubs/StateOfTheInterconnection/Pages/Western-Interconnection.aspx The Eastern Connection serves most of the US and Canada. The smaller interconnections each serve much smaller populations. So, a blackout that affects 60M could affect one or more of the interconnections.
@das_gug also laid out a pretty comprehensive list of possible causes; I'd add war and stupid glitches (see 1965)/squirrel disasters. @404_NOT_FOUND linked to the Nature paper with the best current estimate of how likely a Carrington-level solar storm is, but it should be noted that a much lesser solar storm, such as that seen in 2003, could be sufficient to cause a blackout affecting >60M. From the abstract of the Nature paper: "in particular, the probability of occurrence on the next decade of an extreme event of a magnitude comparable or larger than the well-known Carrington event of 1859 is explored, and estimated to be between 0.46% and 1.88% (with a 95% confidence)." So if we say the chance of a Carrington-level event in the next decade is about 1.2%, and perhaps that of a 2003-level event might be approximately 10% (guessing here), then the chance of a solar storm like that in 2003 might be a bit less than 1% in the forecast period. The grid has been a bit hardened since then, but not nearly enough.
If we add in the risks of cyberattacks, natural disasters, war, physical terrorist attacks, and stupid glitches/squirrel disasters, I'd put the risk at about 2%. I also put my forecast above the historical average because severe weather events are becoming more severe and/or common, the grid has a lot of problems and there are simply more and more people interacting with it, and I think that all of those things raise the risk.