Kaleidos – A portable nuclear microreactor that replaces diesel generators
173 comments
·July 31, 2025acidburnNSA
no_wizard
Wonder if much of the world didn't turn away from nuclear power they way they did since the 1960s, if we wouldn't have solved alot of problems like these already given research was stagnant (relative to other research in power generation) for a very very long time.
acidburnNSA
It'd be a much different field if we had kept it up. I spend a lot of time in nuclear archival material, and facilities like CANEL in Middletown CT absolutely blow my mind. They had hundreds of people working on crazy reactor technologies. They were flowing white-hot lithium metal at 100 mph. But yeah we gave all that up. My friend wrote a pretty good article about this not long ago https://www.ans.org/news/2025-05-08/article-6961/hightempera...
binary132
In order for this to happen, making websites and mobile apps is going to have to get a lot less lucrative.
ViewTrick1002
I have a hard time seeing how communities that have trouble keeping the skills necessary to operate diesel generators will be able to switch to nuclear reactors.
https://www.spitsbergen-svalbard.com/2024/04/09/longyearbyen...
AnthonyMouse
I don't really get the "make it small enough to fit on a truck" thing. The main impediment for nuclear is cost, and then being able to build reactors on an assembly line would be a significant advantage. But how much of that advantage is retained if the product comes on more than one truck and the thing that comes is the reactor, the fuel and the turbines whereas the concrete gets poured on-site? It seems like that should get you nearly all of the cost savings from mass production but then you get a full-sized reactor that can power a city instead of something that can only replace a diesel generator.
ortusdux
> I hope they can find a way to bring fuel costs way down.
I've spoke with some researchers and investors working on seawater uranium extraction and left quite optimistic.
philipkglass
Extracting uranium from seawater gives you natural uranium, which needs to be enriched for use in most power reactors. The reactor under discussion here needs higher uranium enrichment and more expensive fuel fabrication operations than common power reactors. Developing uranium extraction from seawater is a good long-term insurance plan for uranium availability, but it's not going to help this reactor get its fuel costs down.
cyberax
> On the plus side, it's super robust and can minimize the need for other safety systems.
Can it survive 20 kilos of TNT planted by a terrorist?
acidburnNSA
If they radiation shield it properly, I'd like to think so. That won't do anything to 8 ft of concrete plus 4" of tungsten.
Plus the fuel form holds in a lot of the fission products even when scattered around. It may overheat and release volatile fission products but I don't think it would be a widespread disaster no matter what.
Atotalnoob
Could you reduce the amount of concrete by increasing the amount of tungsten?
cyberax
Well, how about a 500 kg shaped charge?
And you can just remove the control rods and wait for it to melt on its own (meltdown-proof fuel isn't actually, well, proof). You'll get a nice contamination with volatiles (cesium and iodine), for the bonus points you can wait until the end of the fuel campaign to maximize the amount of transuranics.
I just don't think this is a viable option, except for very niche scenarios.
whamlastxmas
Can a skyscraper?
credit_guy
Pro tip: if you want to check if a nuclear reactor design is vaporware or has real legs, you check their application with the NRC (nuclear regulatory commission). It turns out that the NRC does have a page for Kaleidos [1]. You can even go and see what documents Kaleidos submitted this month. They submitted a request to be exempted from some regulation called "10 CFR 55", which they feel is not applicable to them. I have no clue if that is the case or not, but at least they seem to be in fairly frequent contact with the NRC, and that's good news.
[1] https://www.nrc.gov/reactors/new-reactors/advanced/who-were-...
master_crab
10 CFR 55 covers operator licenses. The regulation does allow exemptions in certain cases where licensing isn’t necessary.
SoftTalker
Am I right that 1MW of solar generation would only take about a football field worth of panels? Of course that doesn't account for battery or other storage for nighttime, etc. but seems like it would be far cheaper and far less regulatory issues unless you really needed that much power generation in a very small footprint.
generalizations
"Only"
Looks like a giant part of the value is that it can be shipped in, dropped on the ground on site, turned on overnight, and it only takes up the footprint of a shipping container.
If you have 24 hrs to find an empty football field within a powercable's distance of what you're trying to power, and then fill it with solar panels and batteries, you're gonna have a bad day.
nehal3m
If you ship in a stack of panels, inverters and cables, sure. But maybe you could be a little smarter about it, like a container with all the electronics (inverters, batteries, management) and a bunch of folded, pre-cabled panels that you can pull out across a field. If you bring a couple of those covering a field in a few hours shouldn't be that hard and could be ready for use instantly provided the batteries are charged at delivery.
generalizations
You also have to make sure there's an empty football field onsite. That's a much harder ask than dropping a shipping container in a couple of parking spots.
djfobbz
Are you're seriously comparing a few containers of fold-out solar panels and batteries to a portable nuclear reactor?
Let's do the math: To match even a 1MW reactor, you'd need 2,000+ panels, inverters, batteries, mounting, and approx. 120-150 man-days of labor...and that's with pre-cabled gear. You're still looking at 8+ containers, a full crew, and a full 10-14 days to deploy, not "a few hours."
A nuclear microreactor doesn't need 54,000sqft of land or weather-dependent storage. Nice idea for a solar camp but not a replacement for a compact nuclear source.
denkmoon
the sun doesn't shine at the poles 6 months of the year. it usually doesn't shine underground or deep underwater. etc.
In places solar panels make sense they would certainly be used, but that's not everywhere.
supportengineer
I can imagine a Falcon 9 dropping off an automated package that unrolls/inflates the solar panels. Security would be provided by even more drones.
generalizations
Still need a place to unroll them. Unless you have a football field of empty space in that falcon 9, too.
esseph
You're not accounting for location at all.
Nor is that generating electricity at night.
Plus battery storage.
And it's closer to 4-6x football fields if you did it in say, San Francisco. 4-5x football fields in Kansas City. 6-8x football fields in Chicago. Again, plus battery storage.
hinkley
How many power problems would be covered by making a battery the size of this device
actinium226
Depends entirely on how you charge the battery.
quickthrowman
Not many, it’d be roughly 1MWH of storage, based on looking at battery storage in half size conexes that are currently available for sale. A 1MW reactor or diesel genset can put out 1MWH every single hour for as long as they have fuel.
esseph
zero
daemonologist
It depends on where you're at, but for a sunny place yes; somewhere like London a panel can harvest ~100 W/m^2 (0.5 MW for a football field with 100% panel coverage) averaged over the whole year, while in Arizona it's more like 230 W/m^2 (1.2 MW for a football field). NREL has some great insolation maps here: https://www.nrel.gov/gis/solar-resource-maps
For a permanent installation I would agree that solar would usually make more sense, but the mini reactor might be better in scenarios where it's replacing a diesel generator - emergencies, temporary events, confined spaces, etc.
asdfman123
It requires space, setup time, and then there's the intermittency issues. You'd need enough batteries to store, what, 12 MWh? 20? More if you're accounting for cloudy days?
People just want a compact solution to generate power, not a whole separate project.
jonplackett
Wouldn’t the point be that it works at night too?
quickthrowman
“Only” a football field is a lot of space. A 1MW diesel genset on a trailer is about 30’ long by 8’ wide by 10’ tall, which is 0.4% of a football field.
trklausss
I'm skeptical, not because it can't be achieved, but because it's not that practical.
Diesel generators are "great" because diesel doesn't evaporate. You can have it there for years, and with good design, it just springs up the next day.
This nuclear reactor has to be connected for fleet monitoring if you want to operate it. Which excludes it from many real life scenarios where diesel generators are used.
Maybe for remote locations where constant power is needed (Antarctica and such), but I see their uses being very limited.
p1mrx
If I Google "diesel shelf life", the most common answer is 12 months. Do you have a better source? Propane probably makes more sense for fuel that needs to sit around for years.
Do you know the shelf life of TRISO fuel? I imagine it doesn't matter because it would be very expensive to build a reactor and not switch it on.
SR2Z
Diesel will degrade with exposure to oxygen, but a diesel engine can burn pretty much any flammable liquid that you can meter out. It really comes down to the engine itself and if it can handle less-than-perfect fuel.
whatever1
It can even burn its own lubricant oil and die in a screaming runaway fashion!
lb1lf
Anecdotally, I came across a large (for a single user) quantity of diesel 9 years ago. (Nothing exotic - a company went titsup and I was the only one both bidding for and capable of removing the diesel from their premises within an acceptable time frame; I got approx 80% off the pump price at the time.)
I still run my tractor and Land Cruiser off the stuff; the tractor had an outing today. Granted, neither of those engines are very particular about the fuel they are given, but still...
(Water drained off every few months, also a biocide is added to keep the diesel gunk at bay.)
AngryData
I mean if you trying to run that fuel in a performance application where you are pushing the fuel to its absolute limit, it might be bad, but most diesel engines can be run on nearly any burnable oil, you just get less power out and a bit dirtier burn.
They give similar specs ideals about gasoline fuel going bad in 3-6 months, and yet 95% of gasoline engines will still run 2 year old fuel fine because they aren't pushing compression ratios to the absolute possible limit, and half of the performance engines that do push limits these days have adaptable computer controlled compression and sensors which will figure out how much it can push the fuel.
If I put 5 year old diesel fuel into any regular diesel motor or generator or vehicle and it didn't start up, I would be extremely surprised, and be most worried that the fuel either wasn't diesel fuel to start with or had a wide open hole in the container that a bunch of rain water drained down into.
That said, if I had some kind of tuned up diesel motor that I was trying to push 800+ HP out of, I probably wouldn't use year old diesel fuel just in case. High performance motors like that are already straddling the line between working great and catastrophic failure and using old potentially bad fuel only adds to it.
actinium226
> This nuclear reactor has to be connected for fleet monitoring if you want to operate it. Which excludes it from many real life scenarios where diesel generators are used.
I don't understand this sentence, why does connection to fleet monitoring preclude using this microreactor as opposed to a diesel generator? Can't you just hook a starlink up to it, and program it to shut down in the event of prolonged comms loss?
garte
Adding an Elon company to the mix might make the reaction unstable.
rich_sasha
I don't have any first hand experience with diesel generators, but I saw three cases where power was lost and diesel backup was switched on. In two of these three cases, the generator failed (once didn't start, the other time it ran for 30 mins). In both cases it was in scenarios where I'd imagine reasonable care and maintenance were applied.
cyberax
> Diesel generators are "great" because diesel doesn't evaporate.
LOL, no. I see, you have never worked with large diesels meant for backup.
If you just leave diesel fuel alone, then over time (6-9 months) the residual water separates at the bottom of the tank. And then various microbial life springs into action, happily living off all of that free energy. While there's some dissolved oxygen, it will happily use it to oxidize the fuel. But even without oxygen, the bugs will try to live off energy produced by polymerization of unsaturated hydrocarbons.
Polymerization == gunk that clogs up your fuel filters.
So you have to periodically clean up diesel fuel by removing water and filtering the gunk out. It's called "fuel polishing". Large diesels will have fixed systems, for smaller diesels, sometimes mobile systems are used like these: https://fueltecsystems.com/equipment/pneumatic-systems-2/
mikewarot
At this point, I'll assume that the relevant US regulatory agencies are competent, and skip the safety issues, etc.
What does it cost?
How much power can it deliver?
So what's the equivalent $/KWh?jgeada
That used to be true, not so sure my trust in our institutions is high these days. Seems a few million $ donation to the right people can make all regulations just vanish.
acidburnNSA
Yeah there's currently a large push to dismantle the Nuclear Regulatory Commission and give reactor regulation authorities to the states.
jabl
For all the faults of the NRC (real or imaginary), this still strikes me as an extremely bad idea.
dylan604
That's pretty much always been the case. It's just much more flagrant and in the open now.
brink
> That used to be true
No, that used to be believed to be true. We're just seeing the curtain come down.
The food pyramid, the CIA's "war on drugs" in South America, the wars with Iraq, Libya.. Just to name a few. Why do we pretend like bribery and corruption is this new thing?
mrtesthah
That sounds like an excuse to abide the open corruption that's occurring under the right-wing regime in the US today.
Kon5ole
From the FAQ:
> The plan is for the small amount of spent fuel [..] that comes out of our reactors at the end of their duty cycle to only be temporarily stored on-site until a federal repository or interim storage solution becomes available.
>The NRC is legally required to ensure that, in the event of bankruptcy that the site is secure and the spent fuel remains safely managed before ultimately transferring over to the U.S. Department of Energy's (DOE) possession
So basically like with all other nuclear, the tricky parts with unknown costs are handed over to future taxpayers. These costs are thereby removed from the estimates in Kaleidos business plan making it seem like it's a viable business, which is isn't otherwise.
Meanwhile the founders get rich today, since immense amounts of money has to be presented upfront, before any energy is produced at all.
The obvious replacement for diesel generators today is solar+batteries, which is evident in many countries that used to rely on small scale diesel generators already.
Animats
Here's a related company, Ultra Safe Nuclear, making TRISO fuel units.[1] They went bankrupt in April 2025.[2] They at least got as far as making fuel units, although I suspect the video shows dummies being made, because they are not taking enough precautions for handling enriched uranium.
[1] https://www.youtube.com/watch?v=uR7VDqUbaCg
[2] https://nuclear-news.net/2025/04/04/update-on-the-bankruptcy...
gmueckl
I'm not filled with optimism about this concept. Let's work backwards from crash safety (say a reactor on a truck getting t-boned by a freight train). The radioactive material needs to be held in an armored containment to avoid release. That would have to be roughly comparable to CASTOR containers in terms of its resilience. But these containers have limited capability of passive thermal energy dissipation (Google finds models that handle 10kW to 45kW thermal power generated in the interior). This would be approximately the ceiling for the direct thermal power output that is still reduced by limited efficiency of heat-to-power conversion.
This is admittedly napkin math, but it should be good enough to set expectations.
jillesvangurp
You are thinking accidents. I think we need to be thinking deliberate attempts to compromise these things and all the security measures needed to mitigate against that. And most importantly, the cost associated with that. Which comes on top of already significant cost.
The naive notion of we'll just ship these all over the place by the thousands and it's going to be fine is not going to withstand a lot of critical thinking very long.
gmueckl
I was indeed not thinking about deliberate attacks. But that doesn't change the result much as I assume that the CASTOR containers that I used as reference are designed to withstand all of these worst case scenarios.
conorcleary
Unfortunately, this chain of messages underscores the importance of keeping humans in check with eachother, and the machine.
altcognito
Hundreds of thousands if we're talking about meeting growing demand. Which doesn't give me a lot of hope for solar either though.
jillesvangurp
Solar already runs circles around nuclear when it comes to cost. Typically combined with batteries. The combination is popular and cost effective.
With nuclear it's all rosy and optimistic. But also almost 100% hypothetical. And the industry has a piss poor record delivering on its promises. 200-300% cost overruns are routine.
We won't see more than a handful small small nuclear reactors for years to come. We might get to hundreds by the 2040s or so. Maybe growing to thousands or even tens of thousands by the 2050s under the optimistic scenario.
Most of these things have the power output of a handful of wind turbines, of which we have close to half a million right now around the world with more coming online all the time. The challenge here is that wind turbines are just stupidly cheap and scalable at this point and still getting better.
SMRs will remain very niche for a long time even if they do get their cost levels under control. Which is a big if. Thousands of these things would barely move the needle in terms of power output. Essentially all of the expected growth in electricity demand for the next few decades is going to be met by wind and solar supported by batteries.
ranie93
Could this be associated to a supposed recent State Department approval?
“I just approved a program to deploy small modular nuclear reactors built in the United States to an allied country to help with their sort of energy infrastructure.”
“Which allied country would that be?”
“I can't tell you. It's not public yet.”
From Interesting Times with Ross Douthat: The DOGE Alum Asking if Foreign Aid Is America’s Problem, Jul 31, 2025
linuxguy2
How much does it cost? Would love to buy one with the HOA and run our own micro-grid while exporting electricity to the local utility.
unglaublich
When HOA gets too powerful.
GuinansEyebrows
This is bringing “My HOA went nuclear on me!” to a whole nother level.
bostonwalker
Not in my back yard!!
KaiserPro
From the headline I was assuming it was a tiny 20kw job.
But it being a 1.9mw(thermal) makes sense.
I wonder what the support requirements are, like how do you yeet the heat to make it efficient?
Also containing super heated helium seems hard for any length of time. I wonder what the operating lifespan is.
t0mas88
They say it needs to be refueled after 5 years and that it can be done 4 times for a total lifespan of 20 years.
Reason077
1.9 MWt still seems like a huge amount of energy/heat for something that fits on a truck and is supposedly air-cooled (they claim no water is required).
Where does all that heat go?! They must have some very impressive fans.
tralarpa
Yes, that's crazy. They say up to 1 MW electric which would mean (33% efficiency) 2 MW of heat to get rid of with air cooling. Later they mention facility heating which sounds more realistic, I guess?
KaiserPro
I mean its not that much different from a diesel generator, they are around 30% efficient, so they'd also be kicking out the same amount of heat?
https://www.generatorsindustrial.com/products/1mw-diesel-gen... has a simple radiator.
but then the heat profile is different I suppose, and the efficiency doesn't depend on being able to shed heat.
colechristensen
Heat transfer has the lovely property of scaling nonlinearly by temperature difference. You need a lot of big fans to cool your CPU from 75C on the die to a 25C room, instead of a 50C difference these reactors will dump heat at hundreds of degrees C warmer than the local environment.
hagbard_c
Meet EMD DDA40X [1], the most powerful diesel–electric locomotive model ever built on a single frame incorporating two diesel engines with an effective power output of 4920 kW. Given the expected losses in the diesel engines (~40-45% effective, 60-55% waste mostly in the form of heat) and diesel-electric traction system (power generation, traction motors, gearing etc, around 80% effective) which gives a total system efficiency of around 35%. Assuming most of the waste energy ends up as waste heat this ~30m long locomotive (a bit more than two 40ft containers) needs to shed around 9 MW of waste heat or about 4 MW per 40ft standard container length.
codezero
[dead]
ACV001
A startup where most of the money were spent on that animation on the website.
dylan604
You'd be surprised at what the CAD software can do now in 3D renders. You have to design the thing in CAD anyways, so it's not like the 3D team had to model it from scratch. You could probably just do this with a request on Fivr. These aren't your parents 3D prices any more
This is a passionate team working on a very hard problem. They have guts and skills. I've always loved microreactors for fringe remote power where people are willing to pay 20x more than normal diesel generator prices. Like Antarctica, remote bases, the moon etc.
Trying to make microreactors cheap is super hard. We've obviously tried it many times, the most relevant being the truck-mounted military microreactor ML-1 (the only closed-cycle direct gas turbine reactor ever operated) https://en.wikipedia.org/wiki/ML-1.
Shielding is hard. Even a small reactor this size needs like 8 ft. of high density concrete on all sides, or equivalent, plus 4-6" of a heavy metal like tungsten to take down the gammas. You can't just put it underground because the neutrons activate the dirt. Driving it off afterwards is borderline impossible because you generally have to put the spent fuel in robust canisters that can handle collisions, rollovers, and RPG attacks.
But the hardest part is fuel cost. This reactor uses medium-enriched ('HALEU') fuel, which is super expensive, and then it packages it into TRISO form, which is about 100x more expensive to fabricate than regular UO₂ fuel. On the plus side, it's super robust and can minimize the need for other safety systems. Those prices could both go down, conceivably, but the fab process is pretty intricate, and it's hard to bring down enrichment costs. In my analysis, the fuel cost alone nearly makes this kind of reactor uncompetitive with a diesel generator in almost all applications. So even if the reactor is free (because you build it on an assembly line?), you're still out of luck.
Then there's thermal strain. When you're a small reactor you have big gradients. This bends things. Neutrons make it worse. Then you have a tiny box with electronics in it getting absolutely hammered by neutron dose. That does bad things too.
I hope they can find a way to bring fuel costs way down. I really like the people at this company, and I really like nuclear power and want to see it used in many new applications. I just don't quite see the path yet.