A few ways
1: Selection bias - how can be we sure IM isn't showing us data selected to be significantly different? There are various lifting and causings of restrictions that aren't listed.
2: Regional differences - are Arizona, California, and Nevada actually comparable? Are the lifting of restrictions exactly the same?
3: Presentation of data - All of the relevant data is in the bottom right corner of this image and it's excessively zoomed out in a way which will inherently make the differences seem more minute.
4: The person who made this image has not actually done any math to show a significant difference or lack thereof. California, still having a mask mandate, seems to have significantly less cases than Nevada+Arizona which both don't after June 1st, and Nevada specifically seems to have rising cases. Is the actual premise "Virtually no difference in COVID case rates" actually true?
5: Is the information in the picture even accurate or fully accurate? Arizona maintained mask mandates for government buildings and public transit, what opened in Nevada is clark county (most of Nevada) but not Nevada itself. There is unsourced claims.
It's mostly some coomer putting some lines next to each other and going "HURRFF THIS PROVES NO LOCKDOWN WORK" with delusions of intelligence and self-importance who is more concerned with OWNING PEOPLE than spending the time studying to learn basic statistical skills required to contribute more than some annoying peabrain contribution. I've long been sceptical of full lockdowns and now mostly believe that they have a mildly positive effect which is outweighed by their downsides and they're not justified on a cost/benefit bias. Further, we could have determined this was the case by mid-2020.
Here's a paper making this argument:
https://www.sfu.ca/~allen/LockdownReport.pdf 
