I would love to see a source to back up the 50% loss of power claim as well.
It's utter BS, of course. Total losses from the primary windings of the first transformer at the wind generator in west Texas to the load at a customer site on the gulf coast is under 10%.
But power has to come from somewhere and we can't burn our way out of the problem. wind and hydroelectric will always play a role but as with all renewable sources they aren't constantly available everywhere. Barring someone figuring out how to do fusion, it is likely solar with storage is the way out. the storage part of that is what has to get worked out along with increased efficiency of solar panels (which have come a long way with a lot of room to grow).
Even with BATTERY storage, at the well financial "learning curves" of solar, wind, and batteries, overbuilding the wind & solar by 3x-4x produces enough energy to get by with only 1-4 days worth of battery storage, and by 2030 would be cheaper on a lifecycle basis than all thermal power generation (gas, coal, nuclear), even cheaper than the operational costs of existing powerplants. (Solar + 4 hours of battery is already cheaper than simply maintaining & running a gas peaker in Texas or California.)
Traditional thermal electricity sources are a slowly evolving infrastructure, with very low (if any) learning curves due to the low numbers of build-out. Solar wind & batteries are all technology, scalable, with very rapid learning curves as manufacturing volumes scale. Any real technology breakthroughs (more likely in battery & solar than wind) would change the mix, and shorten the time horizon. Small modular nuclear might have had a shot at being competitive if it were in it's current state of readiness back in 1990, but it's unlikely to be competitive against wind/solar/battery by the time the first units are going on line.
The optimal wind/solar/battery mix would vary by location, but Tony Seba's group has modeled it for California, Texas, and New England looking at 40 years of weather data, modeling it with conservative learning curves. and it works even in less-sunny not so windy (on shore, anyway) New England to come up with the above graphic.
The other nice thing about Seba's wind/solar/battery model is the fact that for only ~20% more capital up front something like 300% more "extra output" than is needed for current energy uses gets created at effectively zero marginal cost. This nearly-free and gia-normous energy output means the entire energy market (all types and uses) is going to be severely disrupted well before 2050, and that is going to happen independent of government policies, subsidies, etc. The sheer economics of it will be the driver.
Most people, including policy planners or corporate CEOs have a hard time intuiting exponential growth curves and tend to think linearly, but learning curves are real, and quite non-linear. As markets saturate it becomes a logistic curve, but we're currently at the rapidly thickening wedge a the thin edge of the curve on batteries solar & wind, where it's more similar to an exponential curve:
With every doubling of solar production volume the cost drops by about 20%. With wind power it's maybe half that, but with batteries it's even faster. Even in dirty-air lousy pollution controls China the lifecycle cost of wind & solar are already at parity with dirty-coal (without factoring in the health care or climate cost externalities of dirty coal.)
The investment banker analysts at Lazard for the past decade or so have produced annual updates on the levelized cost of new electricity generation sources in the US, and now the levelized cost of storage looking at different use cases, and testing for sensitivity to subsidy, etc. It's important for those in the energy industry to keep track of that stuff when deciding whether to fix or rebuild damaged/aging powerplants or pipelines.
Spending a lot of money for new infrastructure, be it pipelines or powerplants that will take 25+ years to break even is VERY risky in the face of the technology learning curves of wind & solar & batteries. Ratepayers in GA may rue the day they fuel up the shiny-new Vogtle nukes, since they'll be stuck with the decommissioning costs on a plant that will never even be able to pay off the costs of construction. SC did the right thing by halting the construction of Vogtle's sister plants for SCANA.
In free-market ERCOT territory this disruption may happen even sooner than in the more regulated parts of the US.