Dr. Jason Hickel has—I think—a pretty fair criticism of these graphs [1], that I'll try to summarize here:
1. Treating "poverty" as a binary "yes/no" thing, where having > $1.90/day suddenly alleviates you from poverty seems like a bit of a stretch. $1.90/day is probably way too low a number to cover adequate nutrition, shelter, and basic human needs.
2. While the percentage of people in poverty by this measure has decreased, the total number of people is probably more important. And that number has actually increased in the past 40 years.
3. They didn't start collecting these numbers since 1981. All numbers prior to 1981 are highly dubious.
4. It ignores the effects of colonialism. Many people went from living lives with little need for money to one totally dependent on it. Enclosure of what once were public resources has eliminated many non-monetary forms of wealth these people once had.
5. Related to 4, it's not clear that going from a pre-monetary system to a much more capitalist system where money is necessary for virtually everything is an improvement. Especially when that transition was brought about by the threat of violence.
6. Virtually all of this reduction in extreme poverty occurred in China, which makes the notion that this was a triumph of capitalism / neoliberalism kind of silly.
[1] https://www.theguardian.com/commentisfree/2019/jan/29/bill-g...