I agree with most of this. Chinese EVs, power electronics, and solar and wind power generation hardware are in extremely strong positions and have a lot of market share. Economies of scale will make it hard to unseat them. The world will have to build electrical generation capacity at a staggering rate to transition to EVs, a transition that is very likely to happen due to their lower costs and higher efficiency. There is a major risk of backdoors and remote shutdowns.
I do have three quibbles, though.
1. you say:
> Nobody has enough energy to support a full transition to EV's.
Probably what you meant to say is that nobody has enough electrical power generation to support that transition. Conventional, full-sized electric vehicles will, for the most part, be displacing internal-combustion-engine vehicles which consume about three times as much energy. That energy consumption will evaporate. Transport is about 27% of US energy consumption and about 15% of world energy consumption; electrifying all of that transport would drop those numbers to 9% and 5% respectively. (Reality will stop somewhat short of that because electrifying long-haul airliners is not in the cards.) So we're talking about a significant drop in energy consumption, not an increase.
But it will be a significant increase in electrical energy consumption. The US, for example, currently generates about 460 gigawatts of electrical energy (using, mostly, much larger amounts of primary energy) and uses about 800 gigawatts of energy for transportation, so we're talking about adding another 270 gigawatts, average, of electrical generation. This will probably be mostly solar, which will require about 1400 gigawatts of nameplate solar power generation, assuming a 20% capacity factor. This amounts to about US$7 billion of investment at the current bloated US equipment prices, which will hopefully come down to more like US$1 billion before all is said and done. Additional transmission and distribution capacity may triple that. The intermittency of solar doesn't bother EVs at all—quite the contrary, those EVs amount to something like 3000 gigawatts of decentralized, dispatchable electrical energy storage, with a total storage capacity on the order of 70 PJ ≈ 20 TWh.
The above assumes that the EVs are exactly as efficient as the ICEs they replace once the energy gets out of the engine, rather than, for example, being lighter or having regenerative braking. I'm only assuming a factor of 3 reduction in energy use due to using 90%-efficient electric motors instead of a 30%-efficient ICE. There are additional interesting effects from EVs being more efficient and able to scale down to the scale of scooters, which we should expect to increase the energy usage of transportation.
2. EVs are only cheap if you can recharge them cheaply. So EVs steamrollering the global auto industry is a scenario preconditioned on the rollout of that electrical generation, transmission, and distribution capacity. Nobody will buy EVs if they expect frequent days-long blackouts or extortionate prices for electrical energy. This may be a spur to the deployment of decentralized solar generation even when it's inefficient. But in any case it seems like a regulatory mechanism that will keep the transition relatively orderly.
3. You say:
> Finally, the renewables power generation infrastructure requires maintenance at roughly modulo 25 years. And, once again, all of that material will come from China.
I have three separate subquibbles with this.
3.1. The 25-year thing is kind of a myth. (I corrected you in https://news.ycombinator.com/item?id=42424748, 19 days ago, the last time you made this false claim, so you already know, or should know, that it is false. In https://news.ycombinator.com/item?id=42539289 you seem to be saying that you don't care whether the things you're saying are true or false, but hopefully I've misunderstood you or you've changed your mind, and at any rate other people presumably do care.)
In more detail: solar panels are subject to significant infant mortality from things like hotspots, breakage, and delamination, but after that, crystalline photovoltaic capacity degradation is a pretty continuous process, typically at about 0.5% per year. Usually this is modeled as a linear decline, which means that you have 87% of the original capacity at 25 years of age, 75% at 50 years, and 50% at 100 years, but of course nobody has panels that old to test with. We do have 50-year-old panels, and they seem to have degraded significantly less than that.
Solar modules typically come with a 25-year warranty because what is being warranted is that they will produce at least 90% of their rated power for that time. (To ensure this, initially they produce a little more than their rated power.)
Some degradation mechanisms, in particular corrosion of copper conductors in the panels by water, catalyzed by acetic acid released from the UV-induced breakdown of EVA glue, as well as browning from that same breakdown, seem to accelerate over time. Others, like milky discoloration of that glue, slow down. It's possible that those that accelerate will dominate, causing panels to fail en masse rather than slowly degrading, especially in humid climates where corrosion is more of a concern. But that won't happen all at once at 25 years; it might happen at 30, 40, or 60 years, or not at all.
The most in-depth paper I've found on this is "Degradation analysis of photovoltaic modules after operating for 22 years. A case study with comparisons," https://www.sciencedirect.com/science/article/pii/S0038092X2... showing lots of different kinds of degradation, with photos. It mostly focuses on a particular case study in Spain, but is careful to explain how circumstances vary elsewhere. "Long-term degradation rate of crystalline silicon PV modules at commercial PV plants: An 82-MWp assessment over 10 years," https://onlinelibrary.wiley.com/doi/full/10.1002/pip.3456 covers 13 different commercial PV plants in Spain and Portugal and analyzes different modes of degradation seen there in somewhat less detail. Finally, "Technology and Climate Trends in PV Module Degradation," https://www.nrel.gov/docs/fy13osti/56690.pdf, is a slide deck from NREL from 02013 that summarizes the degradation rates seen in several thousand studies, including amorphous silicon and other thin-film cells and many degradation studies over more than 20 years.
(Due to US protectionism, First Solar is still making and selling a significant number of thin-film panels there, and those may have different failure characteristics than the monocrystalline silicon panels that are almost universal elsewhere.)
So, probably, for the most part, people will respond to solar generation capacity degradation by adding more panels, not replacing existing ones.
3.2. You seem to be envisioning a scenario where the electrical grid gets built out by more than 50% over the next few years in order to support a transition to EVs—and then the buildout stops, because I suppose nobody can figure out what to do with two or three times as much super-cheap solar energy, once they have electrified their current level of travel?
This doesn't seem like a likely scenario to me. Much more likely is that, as the cost of energy goes down (and energy in the daytime becomes almost completely free) people will find ways to "waste" what today we would consider massive amounts of energy on things we wouldn't consider now, just as we waste billions of CPU cycles and gigabytes of RAM in a way that seemed improbable 20 years ago.
This means that, even if solar panels dropped dead at the 25-year mark instead of gradually degrading, we probably wouldn't see a giant bolus of electrical capacity buildout over the next five years, followed by another one in the years 02050–02055. Instead, we'd see gradually increasing amounts of electrical power generation capacity being built every year over that timespan.
3.3. Even if we did see the collapse you are implicitly predicting in new electrical power generation capacity installations after the grid grew to accommodate new EV and AI demands, followed by a giant replacement effort in the 02050s, what leads you to predict that China will still have the virtual monopoly on solar power it enjoys today? Every industrialized country in the world seems to be currently scrambling for energy sovereignty—some in implausible ways like Mexico's giant bet on petroleum and Argentina's bet on nuclear, but others simply by copying China's success, such as Brazil, Vietnam, India, Thailand, Malaysia, and the US—and some of them will surely be successful. And, as other commenters here pointed out, there are already other makers of power electronics.
Thanks for a thought-provoking comment! I'm interested to hear what you think.