Would love to see comments that actually respond to the article! Everyone here seems to be sharing their general opinions on fusion or a reaction to the article title. The article is actually very detailed!
Agreed, I think it is a great article. Personally, I'm most inclined to agree with the "bear case" summarized at the bottom of the article:
1. I think it's highly likely that, at some point (and probably in the not-too distant future with ITER and DEMO), humans will produce fusion reactors that are self-sustaining with net positive output (ignition).
2. Given the inherent complexity and difficulty with fusion, I think it's unlikely to ever be cost competitive with simpler power generation technologies.
3. Perhaps most importantly, at least in the next human lifetime, I think fusion will be largely irrelevant in the quest for carbon-free power generation, mostly as a result of the timescales required. I think right now that everyone pretty much understands that the problem with decarbonization tech is base load and storage: carbon-free renewables like solar and wind are already the cheapest form of energy generation, but they're intermittent and unreliable. Given the latest climate data, I think it's imperative that we develop base load and storage tech in the text 10/20/50 years, or we're REALLY f'd. But I haven't heard even fusion's biggest boosters say that fusion will be ready for large scale deployment in those timeframes, at cost-competitive economics. Point being, by the time I think fusion might be ready for large scale deployment, we better have already solved the base load and storage problems with cheaper tech.
Personally, I don't see fusion as relevant on a time scale that will impact the climate crisis in any significant way, but it could be a major asset in taking us from net zero emissions to preindustrial carbon levels. It could also simply just be a way to achieve cheap and abundant energy in general, irrespective of climate issues.
But as you said, of course it's an open question whether fusion will actually be more economical at scale. It could conceivably turn out to be a day late and a dollar short on Earth, and more relevant elsewhere.
I'm keeping my eye on Helion (the one Sam Altman invested in). Seems they're making some pretty big claims, and have already entered into contracts to deliver to customers by 2030 or face financial penalties. Could still be vaporware, but they seem to think they have something.
"Base load" doesn't really enter into it: you just need storage to make renewables make sense by supplying when generation doesn't meet demand and recharging when demand is less than supply. (or, if you have a cheap "base load" you use storage to even out the peaks and troughs in the same way). Base load is just a somewhat arbitrary line drawn across the lowest point of demand, it's not an actual requirement of the grid like supply and demand always matching.
Base load was a useful concept when steady output power plants with high capital cost produced energy at the lowest cost. It then made sense to use them as much as possible -- to handle the "base load" -- and cover the rest with more expensive to operate but lower capital cost sources.
But that's no longer the world we live in. The lowest cost energy is now from intermittent sources, and the optimal grid design will look very different.
Right. Primary power sources will be solar, wind, and batteries. What we need now are peaking plants with low capital cost and high output. High operating cost when running is acceptable because they won't be running much. They're backup generators. Currently, gas turbines can do that, but not much else does.
Nuclear, fission or fusion, is the opposite. It's all fixed cost. A higher fixed cost than solar or wind.
Gas turbines can not do that if we include the cost of pollution. Currently we do not have anything that has low capital cost, high operating cost, overall cheap, and carbon free.
Natural gas just trade pollution, a problem for the future, in return for cheaper energy price today.
There will be times when the wind doesn't blow and the sun doesn't shine. Storage can come to the rescue if this is a few hours, but over much longer time periods will be a problem. It's rare, but not unheard of. Gas will be part of the solution for many years to come.
> Gas will be part of the solution for many years to come.
True. But if we’re talking about rare - maybe once a year or two - dunkelflaute, then the gas power plants aren’t actually going to consume much gas at all. And thus the CO2 emissions will become negligible.
Those gas power plants already exist in most industrial countries. They’ll have made their return on investment so using them for backup will be fairly cheap. Especially in the coming age of robots/ drones for inspection work.
To get to zero emission we can just use biogas or hydrogen. If the consumption is low this will be sustainable and financially viable since the fuel cost will be a very small part of the total operating costs, and making the hydrogen will be very cheap since there will be an abundance of days with excess electricity.
So I don’t see that we will actually have a problem w.r.t. energy storage. Battery storage for short term has already outpaced pumped hydro. The transition to EVs already imply battery production capacity on a scale to handle energy balancing for hours or even days.
Trash burning power plants is a good solution for seasonal demand. Yeah, reuse and recycle first. But eventually when materials degrade we should burn it to avoid landfills.
The industry is starting to say that hydrogen is not going to happen. It's expensive to make even with hydrolysis, explosive, and extremely leaky. That makes transportation and storage really difficult. Also, I think it contributes to warming when enough leaks.
Hydrogen as fuel won't work, yes; but generating it and immediately making hydrocarbons or ammonia or other chemical feedstocks with it is perfectly feasible. It's already being done. Just not with electrolysis.
Yes. On the super long term - wind doesn't blow, sun doesn't shine, oil doesn't flow, coal doesn't mine. Then the feasibility of nucular is a no brainer. That's all that's left besides renewables, nucular fishin. And when we run of of nucular fishin, and that day will come too... Fusion will be all that's left of the non-renewables. And it will be used surely, provided it's a net-positive delta energy.
Plenty of wars have been fought already over oil and gas, and it’s waay easier to redirect tankers when a geopolitical player starts trying to squeeze someone hard. Note however that pipelines already have been caught in the middle of geopolitical drama.
And oil and gas are energy dense enough (and storing it is cheap/easy enough), it’s possible and common to stockpile months to years worth of supplies. All it takes is some big metal tanks.
Not sure how that is going to play out when it takes a decade plus to build a HVDC line, that line is in a fixed position (so easy to sabotage/destroy) and it’s orders of magnitude more expensive to store electricity - so most places will be lucky to even have a couple of days worth of storage.
I’m honestly not sure which will be more dramatic looking if someone bombs it though.
For somewhere with constant good insolation and low winter energy needs (like the Australian Outback, for instance), not likely to be a problem. Australia has never been meaningfully invaded or bombed either. So centralization is likely not a huge concern for them.
For somewhere with peak energy needs that coincide with minimal insolation (and often wind!), like long dark winters? And temps that can easily result in people freezing to death? And that has a history of conflict with neighbors?
I find that fact that the UK would depend on another country for power generation in a serious way really really dumb. There is no other way of putting it. If we did that we might as well sell off our armed forces and declare global peace unilaterally, that is how naive that option is. It would make the UK, or any country that did that, incredibly vulnerable. We cannot ever lose power.
The UK imports 5 Twh of electricity from France as it is so I guess you should start up the auctions.
After failing to successfully launch a Trident missile for 8 years it probably wouldn't change much.
You've only managed to send to Ukraine what the US spends on parks.
Honestly at some point you all need to accept that you really aren't a meaningful player in geopolitics anymore and focus on getting GDP per capital higher than that of Americas poorest state.
Seriously you can't afford to not dump feces in your waterways, what other option do you have.
Uncle Sam and the Polish will keep you safe. The only threat you need to worry about is the machete gangs you keep importing to keep your Pret delivery under 3 quid.
It isn't FUD. This is something the system operators and utilities are taking very seriously and talk about daily.
We have decarbonized a lot in the US already, but we still have a long ways to go. A future of just solar, wind, and storage is still a very long ways off. We'd need a lot of load that is responsive to price and we're just not there yet. That's unattractive politically. People are fine with renewables until all those fixed costs creep into their bills and they're told that they have to go to dynamic pricing to make you hyper conserve electricity during the times you want to use it the most.
Yes, we'll eventually get there, but I strongly believe that gas will still be a big player as a backstop/reserve...maybe with carbon capture technology which runs all out during times of renewable abundance to counter the carbon output of the gas. Will it be practical though?
It’s not clear there would be much longer time periods. Build enough low carbon generation to guarantee a surplus every day and you have a surplus every day.
Anyone claiming we need days worth of storage is inherently making an argument that days worth of storage is cheaper than simply building more generation and generation is really cheap while batteries aren’t.
I often point to this, but it's such a fun and useful site I'll do it again.
What you get is typically (it depends on where you are) producing steady output from solar + wind requires some hours of batteries (typically much less than a day) and usually a fairly large backup of hydrogen (particularly far from the equator). The hydrogen gets burned in combined cycle plants with a mediocre round trip efficiency; most of the stored energy goes through batteries and back again at high round trip efficiency.
So, yes, one does need days worth of storage, but it's not batteries, it's a rainy day hydrogen account where storage capacity is very cheap.
According to several studies you need up to 2 weeks of storage, not just 24hrs, because checking on the long term, every decade or 2 you get 2 weeks of clouds or 2 weeks of no wind, etc....
AND not OR. To actually need 2 weeks of storage you would need both 0 output in solar for 2 weeks and 0 wind for two weeks and 0 output from hydro. That doesn’t happen.
You still get solar power on cloudy days, it just takes more panels to generate some specific level of power.
The goal is to minimize X$ for generation + Y$ for storage while guaranteeing sufficient supply. Any study approaching things from any other set of assumptions is going to give you nonsensical answers.
There are counter arguments to this. Clouds and wind are local weather patterns. We can use cables to move power around between areas with and without clouds. Moving power over large distances has gotten more feasible with high voltage direct current (HVDC) cables. There are a few projects in the works to move power from e.g. Australia to Singapore and Morocco to the UK. And there are already cables moving power between e.g. Canada and California, Norway and various countries in Europe, etc. More cables means more resilience in the grid. Continent wide absence of wind and solar generation is not a thing that happens a lot. Certainly not for weeks.
Another point here is that demand shaping is an effective way to deal with fluctuating supply of power. By creating financial incentives, you can get energy consumers to scale up or down their consumption of power. Night tariffs are still common in places with a lot of static generation, for example. With solar generation now being so common, we even get occasional negative energy rates in some places where the static generation can't be scaled down.
Batteries and cables are a key enabler for demand shaping. Also, the time windows that energy gets sold for are getting shorter. It used to be that you'd buy x amounts of mwh for some price for hours. It's now getting down to minutes. That means grids can respond more rapidly to fluctuations in supply and demand. And of course it creates incentives for companies to invest in being able to scale up or down their energy consumption from the grid and benefit from these price fluctuations. For example by having batteries and using their roofs for solar generation.
Base load is of course a very flimsy concept and the discussions about it tend to be very hand wavy and rarely cite specific numbers in GW needed. Because as soon as you do that, you can talk solutions: cables, storage, more solar (it always generates some power), etc. And cost.
Hysterical assertions that we need to spend double digit percentages of GDP on things like nuclear or fusion kind of fall over when you apply some rationality to that. How much power for how much $? Maybe do something less mad and cheaper instead. Build some cables. Add some off shore wind. Much cheaper, faster, and way less risky.
Of course the reality is that we still have plenty of base load for the foreseeable future. That's why the vast increases in wind and solar generation, which are now the dominant source of power in a growing number of places, isn't really causing any outages or rolling blackouts. Whatever amount of base load we need, apparently it's way less than we currently have because we have been removing lots of it from the grid.
Given the sheer thermodynamic effect of the earth's rotation I have to ask what would it even take to get 2 weeks without wind? It would probably take at least globe spanning superconductors or very careful heating of the world to ensure constant temperatures in order to not have wind from the temperature cycling. That is 'Dyson sphere builders fooling around with their power for laughs' territory.
Also Commonwealth Fusion Systems put out a lot of interesting stuff about how they are going. Even having a look at their YouTube channel is worthwhile.
Metaculus has a lot of stuff on predictions about when fusion will be made viable :
It's detailed, but dated. Most of the graphs stop between 2000 and 2005. Yet the article is from 2024. Somebody took the easy way out and copied and pasted old graphs, probably.
Whatever happened to Lockheed's compact fusion program?[1]
My long run prediction is that ITER will work and everything people on HN kept insisting would work in 5 years will suffer a similar fate.
I'll stake my claim now that I think Helion is probably not going anywhere (just a vibe, but there's something off about their recent big marketing push).
ITER will "work" in the sense of achieving its goals, but that will all ultimately be pointless, as it's a dead end. A device with a volumetric power density 400x lower than a PWR cannot conceivably lead to a competitive energy source.
1. I think it's highly likely that, at some point (and probably in the not-too distant future with ITER and DEMO), humans will produce fusion reactors that are self-sustaining with net positive output (ignition).
2. Given the inherent complexity and difficulty with fusion, I think it's unlikely to ever be cost competitive with simpler power generation technologies.
3. Perhaps most importantly, at least in the next human lifetime, I think fusion will be largely irrelevant in the quest for carbon-free power generation, mostly as a result of the timescales required. I think right now that everyone pretty much understands that the problem with decarbonization tech is base load and storage: carbon-free renewables like solar and wind are already the cheapest form of energy generation, but they're intermittent and unreliable. Given the latest climate data, I think it's imperative that we develop base load and storage tech in the text 10/20/50 years, or we're REALLY f'd. But I haven't heard even fusion's biggest boosters say that fusion will be ready for large scale deployment in those timeframes, at cost-competitive economics. Point being, by the time I think fusion might be ready for large scale deployment, we better have already solved the base load and storage problems with cheaper tech.