Proposed data centre will release 23 Hiroshima atomic bomb worth of energy a day | Hacker News Reader

Proposed data centre will release 23 Hiroshima atomic bomb worth of energy a day

cbc.ca

18 pointsDeepLogin27 comments

How is comparing energy usage of a facility like a data center to a weapon that killed millions of people many decades ago an illuminating comparison?

It is like comparing a banana to a chair.

It's not how much energy is used, it's how much wasted heat energy is dumped into the surrounding environment. Seems like a much more apt comparison in that light. Second paragraph of TFA:

> An analysis of the mammoth data centre proposed by celebrity investor Kevin O’Leary suggests it would release the equivalent of 23 atom bombs of heat waste into the boreal forest of northwestern Alberta every day.

it would be apt if it was "fun fact, despite being incredibly destructive, a nuclear bomb doesnt leave much waste heat" but also a non-sequitur.

a really obvious electricity in, waste heat out is aluminum smelting in quebec. at ~12GWh a day for the biggest ones, this would come out at a similar scale, but make clear that the data center industry in alberta produces far less waste heat than the aluminum industry in quebec, and we should start looking to shutter all those smelters for wasting too much power

bringing up nuclear bombs when the heat output of a nuclear bomb isnt what people are familiar with is intentionally misleading. it also elicits a "cool, this thing can survive 20 nuclear blasts a day?" response which is pretty obviously not intended and isnt true

Still makes no sense. People do not have an intuition about how much heat a nuclear bomb gives off, and comparing an instantaneous explosion versus a energy in a day is really dumb.

It is pretty dumb. But much less dumb.

It's not dumb at all, it's intentionally misleading; as modern "journalists" are shams that rely on misinforming the public over any sort of integrity to make a cheap buck.

Arguably one of the only professions I'd say are a good thing to be entirely replaced. Can't wait.

Not necessarily disagreeing with your main point, but "killed millions of people" is an order of magnitude wrong. More like around 200,000. My point is not that the deaths from those bombs wasn't appalling, it's that exaggerating doesn't help your cause.

> but "killed millions of people" is an order of magnitude wrong. More like around 200,000.

I stand corrected.

It’s inciteful language that garners clicks. Immediately raises alarms for me that the article is going to be heavily biased. Why not relate it to something more benign like powering “y” amount of NYCs or something.

Fun fact: There's about 200 Hiroshima atomic bombs worth of energy in a banana.

And there was vastly more energy potential in the Hiroshima bomb than was actually released - of the 64 kilograms of uranium in the bomb, less than one kilogram underwent fission, and the entire energy of the explosion came from just over half a gram of matter that was converted to energy.

By the above banana logic; How many Hiroshima atomic bombs worth of energy were in the Hiroshima bomb?

By the logic of banana energy potential, you have to go for all of the energy in the entire bomb, not just energy that could be easily released through fission. That's about 4*10^20 Joules vs the 6*10^13 Joules released over Hiroshima.

which, by Cave Johnson logic, gives you about 6.5 million more bombs per bomb!

A toaster also releases 23 Hiroshimas worth of energy over a long enough time frame. Its when it all releases at once that you should be concerned.

"A toaster is just a death ray with a smaller power supply!"

https://www.youtube.com/watch?v=ibOkPx_Ej30

Listen, I really hate waiting for toast in the morning, alright‽

About the Hiroshima bomb, google tells me “Once initiated, the actual nuclear fission process and energy release happened almost instantaneously in a fraction of a microsecond.”

A day has 86,400 seconds, so that’s about 11 orders of magnitude.

That makes this a fairly meaningless comparison.

> 18 gigawatts of heat a day

Dimensional error.

Do they mean 18 GWh per day = 64.8 TJ / day = 750 MW ?

Or is this thing really going to use twice as much power as the largest nuclear power plant that currently exists?

Im convinced this data center project isn't even real, and has been a long running PR stunt to drive outrageous headlines about datacenters so people don't take the real critics seriously, or smaller datacenters seem more reasonable.

it definitely is real. the so big to be outrageous one is being made in utah, by the same guy

23 hb / 24 h = 17.5 GW

Correction: 16.77 GW

That's what the article says, yes. But is it accurate? The confusion of units makes me question the numbers, especially since I can find no other reference of the power usage or generation capacity of the site.

No, it's not, my numbers were off. I did find the original source for analysis

A. my math was off

1 Little Boy = 63 terajoules https://en.wikipedia.org/wiki/Little_Boy

23 Little Boys / 24 hours = 16.77 Gigawatts

B. https://www.theenergymix.com/wp-content/uploads/2026/06/Wond...

"As recently described, the project would draw approximately 9 gigawatts (GW) of power, [...] Power would come from on-site, behind-the-meter natural-gas generation. At realistic generating efficiency, supplying 9 GW of electricity means burning fuel — and therefore rejecting heat — at a continuous rate on the order of 16 to 18 GW"

Strange fact: in terms of heat energy, the hiroshima bomb released far less energy than the fire it ignited. Much of the total destruction seen in footage was the result of the massive instantaneous fire of a woodem city. Had it been a concrete city, things would have looked very different.

A forest fire can release many hiroshimas of energy every day, and canada has more than a few of those each year.

Hiroshima energy equivalents is surely a perfectly and not insensitive way of measuring energy.

New York City uses 1 hiroshima bomb-energy every 3 hours.

I believe we have reached a new threshold in hyperbolic headlines.

Is it a quantum leap level of hyperbolicity?

Are there studies regarding the improvement of productivity with AI?

Or at least economic measurements in terms of money/profits brought by companies/entities who used it?

I mean of course it's going to depend on the field and how it's being used, but there could be a comparison between how much money was spent making AI overall, and how much actual measurable good things it brought.

And I don't mean the amount of customers paying for AI, I mean an actual demonstration that AI is worth doing.

Reply on news.ycombinator.com