I don't agree with the "AI-generated" claim. Gavin Mooney appears to be a real person working with Kaluza, an Australian company which presents itself as:

The Energy Intelligence Platform

An electrified future will be built on data intelligence.

We turn energy complexity into growth opportunity so energy companies can make a cleaner, smarter system work for everyone.

(So a financial conflict of interest, but one I happen to agree with.) I just attribute it to a "shitty, token attempt at sourcing because nobody really checks these things" mindset.

When energy prices went crazy in the UK a while back I heard of some people getting under a year payback times. My energy usage is much lower than theirs so it would take me quite a bit longer though. A lot of the costs are fairly static.

At this point a battery alone might be a better investment. Cheaper install and using off peak rates to charge could drop my per unit costs from 24 to 8. But I think even that would take years to pay for myself. It's also annoying because the grid should already be fucking doing this! Why should I have to do it myself in a setup that is going to be far less efficient in costs than doing it at grid scales with bulk buying of batteries?

The tech exists today, I can buy it.

Economies of scale 

For DYI plug-in small scale solar and meter running backwards (balkonkraftwerk scenario) for 0.3 eur/kWh break even is less than 2 years.

DYI larger/meter not running backwards but with battery buffering it's longer. Anything else requires a licensed electrician, and that does set you back.

The infographic is using 10c/watt as solar panel only. Your 10 year payback is based on tariffs, permits, sales comissions, and a monopoly utility designed to make solar prohibitive. In Australia, payback is about 2 years. But, yes, at utility scale the lack of BS costs make a giant difference. Under $1/watt installed instead of $3+/watt.