For him, it is a political wedge issue meant to drive partisan policy development.
If we are to be hard nosed climate hawks, we need to look at the reality of nuclear, as it exists today, and the reality of our capacity to change our industrial processes. And having done that, after being very bullish on nuclear starts in the 2000s, I do not see how Shellenberger can be taken the least but seriously. He doesn't address the reality of nuclear on the ground, instead focusing on political statements.
https://www.sciencedirect.com/science/article/abs/pii/S03605...
The PV EROI numbers could use an update pretty much everywhere. I have some ballpark-corrected numbers that I keep in my head but the academic literature badly lags industry. You can read a "new" solar EROI paper in 2020 and after dereferencing all its citation pointers find it's still using key numbers from 2008. Thinner cells, lower kerf loss, polysilicon reactor optimization, crystal growing optimization, higher cell efficiency (PERC), bifaciality, and the increasing prevalence of single axis trackers mean that a state of the art solar farm in 2020 has significantly better EROI than the same capacity farm in the same place in 2010. But industry cares about costs, not EROI, and academics who write about EROI seem to lack the will to gather numbers from current industrial practice.
I also occasionally suspect that good solar EROI analysis doesn't happen because people are anchored to previous conclusions. Whether that's nuclear promoters or deep green/"degrowth" types. Both would like to say "there is no alternative" to their visions of the future and a high solar EROI would demonstrate an inconvenient alternative. Or just inattentive grad students and advisors who assume current EROI numbers should be in line with older ones.
The (decade old) source you provided for energy subsidies appears somewhat dishonest, as the date ranges appear cherry picked. Why is it that the source only looks at nuclear subsidies up until 1999, while subsidies to other energy sources are considered through 2009?
The choice of units in this source is also odd. What have the subsidies been in terms of energy output? At its peak, nuclear was producing some 16-20% of US power. I imagine normalizing on a $/MWh basis could paint a different picture.
Looking at a more recent source [0], it appears that recent nuclear subsidies have been marginal and are far outpaced by renewable subsidies. If we consider prior subsidies a sunk cost, it doesn’t necessarily make sense to subsidize a new, less environmentally friendly, source of energy when our best in class is already excellent.
[0] https://blogs.ei.columbia.edu/2019/09/23/energy-subsidies-re...