Small nuclear reactors could solve data centers' sustainable power problem

midian182

Posts: 9,739   +121
Staff member
Forward-looking: Data centers have become a massive industry, but many of them aren't very environmentally friendly due to their huge power demands, which are often met through the burning of fossil fuels. One proposed answer to this problem is for the facilities to use their own sustainable power sources in the form of miniature nuclear reactors.

The Register highlights a report by Omdia analysts Alan Howard and Vladimir Galabov that notes how data center operators have been early adopters of renewable energy—Google and Microsoft recently announced plans to buy more energy to make their data centers greener—but it is still unavailable in many markets. That's where nuclear energy via small modular reactors (SMRs) could come in.

Unlike standard nuclear facilities that produce gigawatts of power, SMRs typically generate 300–500 megawatts (MW) of electric power, though some produce less than 100MW.

Any mention of nuclear reactors tends to bring concerns about disasters such as Chornobyl, Three Mile Island, and Fukushima, but the analysts write that SMRs pose far less risk due to their scale, simple design, and the inherent safety characteristics of the reactor. They also rely on natural circulation, convection, gravity, and self-pressurization.

Additionally, technology similar to SMR is used in 83 nuclear-powered US Navy ships, including 72 submarines, 10 aircraft carriers, and one research vessel, many located adjacent to large population centers, and there has never been a nuclear power incident in the Navy's history.

But despite their benefits, nuclear waste remains a big problem with SMRs, with spent fuel material taking anything from three decades up to 24,000 years to reach safe radiation levels. However, unlike conventional nuclear power plants that require refueling every year or two, SMR vendors are aiming for every three to seven years—some designs are estimated to operate for 40 years without refueling. It's pointed out that nuclear subs only require refueling every 10 or more years, with new cores designed to last 30 to 40 years. The caveat is that SMRs produce 35 times more waste compared to larger reactor designs.

There are currently no SMRs yet in the US, but Russia has two capable of 35MW each sitting on a floating power plant (above) off the arctic coast. New SMRs are now under construction or in the licensing process in Argentina, Canada, China, France, South Korea, and the US.

SMRs will more likely be suited to large data center campuses with capacities of over 100MW rather than individual data centers, though smaller locations could share excess capacity with other industrial plants.

Don't expect this technology to become commonplace anytime soon. It's likely to be another seven years before we see an SMR in the US, and it could be anything from 10 to 15 years before they're powering a data center campus.

Permalink to story.

 
We reprocess it so most of it doesn't become waste. Funny how that's never mentioned in the media.
yeah, I came here to post that and there are safe ways to dispose of nuclear waste, that's something many people don't consider. There are also small reactors designed to run on nuclear waste. So nuclear waste is very recyclable. And something that people don't really consider when talking about levels of radioactivity. The long something stays radioactive the longer it's half-life is. What that means is that it is inherently LESS radioactive
 
We reprocess it so most of it doesn't become waste. Funny how that's never mentioned in the media.
Yeah, but that's too convenient. Can't go green unless we change everything and throw all the money we (aka, the gov't) can to fix the problem!
If green companies aren't getting rich, we're doing something wrong. /s
 
These reactors are not some saver - they are all special purpose ( see above - where cheap renewables aren't available ) - cheap watts is not one of their claims at the moment - do you think Russia cares how efficient their one is. .

We do need small reactors ( Mars base etc ).

These have been promised for decades - development needs to be open source etc .
I'm sure there is a lot of potential in them.

Large scale ones are not a short term fix for global warming as take billions of dollars to build ( may not pass inspection ) - and will not offset carbon footprint to build in near term ie crunch time .
Cost of renewals is coming down.
Other options like geothermal probably just as promising

yes we need better batteries - and storing energy - eg dams ( excess energy in day pumps water up ) .
More efficiency infrastructure and less grid loss.

There are still lots of easy low hanging fruit to reduce energy/CO2 - reducing ship speeds - huge gain

somethings will have a huge effect - like a cheap carbon light concrete for construction.

As for servers - there is a reason why Microsoft etc build near cheap energy - tries to have then underwater for cooling etc

So these are just niche sources - that hopefully with open development - we can master in coming decades for remote bases ( Mars, Moon, Under the Sea , Space exploration etc ) plus source of necessary isotopes for science and medicine
 
This is so dumb. If you would build the big reactors you wouldn't need small reactors. What an amazing idea ....
These reactors are not some saver - they are all special purpose ( see above - where cheap renewables aren't available ) - cheap watts is not one of their claims at the moment - do you think Russia cares how efficient their one is. .

We do need small reactors ( Mars base etc ).

These have been promised for decades - development needs to be open source etc .
I'm sure there is a lot of potential in them.

Large scale ones are not a short term fix for global warming as take billions of dollars to build ( may not pass inspection ) - and will not offset carbon footprint to build in near term ie crunch time .
Cost of renewals is coming down.
Other options like geothermal probably just as promising

yes we need better batteries - and storing energy - eg dams ( excess energy in day pumps water up ) .
More efficiency infrastructure and less grid loss.

There are still lots of easy low hanging fruit to reduce energy/CO2 - reducing ship speeds - huge gain

somethings will have a huge effect - like a cheap carbon light concrete for construction.

As for servers - there is a reason why Microsoft etc build near cheap energy - tries to have then underwater for cooling etc

So these are just niche sources - that hopefully with open development - we can master in coming decades for remote bases ( Mars, Moon, Under the Sea , Space exploration etc ) plus source of necessary isotopes for science and medicine
Nuclear power is the cheapest electricity in the USA, WTF are you talking about in that word salad?
 
Small reactors yes, fission based no. With the stellarator design finally getting the attention it deserves we may see something in the next 5-10 years. Already some impressive results from one group that is planning small scale fusion reactors in the 100-300MW range.
 
We reprocess it so most of it doesn't become waste. Funny how that's never mentioned in the media.

Sure but only 30% is reprocessed and it's expensive. Which owner of a data centre can afford that cost even if we had the capacity to reprocess the huge extra stream of waste that would arise. And also this has to be done in a very timely matter.
 
Ship it and bury it


you still have to pay the costs of decommissioning the plants :they don't just disappear into a hole in the ground!

cleaning up plants is still cheaper than building and maintaining an entire reprocessing system from-scratch, but you still pay ludicrous fees over big-boys plants
 
This is so dumb. If you would build the big reactors you wouldn't need small reactors. What an amazing idea ....

Nuclear power is the cheapest electricity in the USA, WTF are you talking about in that word salad?

Nuclear power is one of the most expensive - especially to build going forward - 10 times more expensive - if it gets off the ground - This is probably ignoring externalities like the taxpayer picking up the bill for the polluters for the next centuries.- Increased cancer near plants and in workers.

There is a reason large scale nuclear power stations haven't been built that much lately in USA - wow first new one in 30 years in Georgia - only cost 30 Billion - if it works :)

That and the coal fire power stations in Georgia - the clean green state - they are expecting their electricity bills to go up to pay for this white elephant.

It's only disinformation right wing groups posting fake stories about how cheap Nuclear power is.
The ones being built in China , India , Russia etc - probably for plutonium and other enriched fuel for bombs .

Nuclear is for 24 power delivery - a back up - solar is way way cheaper - 10 times cheaper .
Quick google search - so do not know current cost of existing plants

Basically in West - unless cost controlled - that 10 billion estimate becomes $30 Billion

Prepared to cut costs, not protect environment , ignore locals - yeah maybe you can build one for 5 Billion - Advantage of Russia and China - Russia can easily produce super cheap gas electricity anyway
 
Nuclear power is one of the most expensive - especially to build going forward - 10 times more expensive - if it gets off the ground.

Why? How much of "the cost" is actual material/work/tech and how much is "something else"?

This is probably ignoring externalities like the taxpayer picking up the bill for the polluters for the next centuries.- Increased cancer near plants and in workers.

Everything has "externalities": coal and pulmonary illnesses, wind and bird killings, solar and global heating... There are ways to control the radiation exposure of people, I would assume "the boss" won't force you to stay after your levels go high. The plants shouldn't be built next to populated areas for the same reason, this sounds like a red herring.

There is a reason large scale nuclear power stations haven't been built that much lately in USA - wow first new one in 30 years in Georgia - only cost 30 Billion - if it works :)

Exactly, why? The ones that are around still work and apparently there's no rush to decommission them. Germany tried it and apparently didn't go well for them (they're holding onto that last one).

It's only disinformation right wing groups posting fake stories about how cheap Nuclear power is.
The ones being built in China , India , Russia etc - probably for plutonium and other enriched fuel for bombs .

"disinformation", aka people saying stuff I don't like. It's a bad move, don't attack the messenger.

Nuclear is for 24 power delivery - a back up - solar is way way cheaper - 10 times cheaper .
Quick google search - so do not know current cost of existing plants

According to your logic: most expensive -> nuclear -> solar -> oil -> coal (I presume). Question: why would anyone build a nuclear plant on an oil producing country?

BTW, you hadn't talk about "the good stuff": you get a highly reliable power source that doesn't burn anything, it doesn't produce "that scary" CO2 and the really scary CO, NOx, etc (or the micro-particles of coal)
 
Back