I keep hearing

“A quantum computer will consume so much power that it'll need a dedicated power plant.”

I'm trying to find literature to back this up. I would love to get an estimate for the power consumption of a qubit and / or quantum gate.

The best I've been able to find is [Ikonen, 2016] which lists this in Appendix A:

Typical powers at the chip are of the order of $P_\pi = 10^{−11} W$, after being generated in the room temperature and attenuated by tens of decibels on their way to roughly 10-mK base temperature. Using only 10 dB of attenuation at the base temperature, the total power dissipation here becomes $P = N_qP_q×10 ≈ 2 mW$. Such power level is much higher than the typical cooling power of 10 µW in state-of-the-art dilution refrigerators at 10 mK.

This paper is a little beyond my physics, but am I correct that the 2 mW figure can be used as a rough estimate of the power consumption per qubit?


migrated from crypto.stackexchange.com Nov 7 '17 at 17:59

This question came from our site for software developers, mathematicians and others interested in cryptography.

  • 5
    $\begingroup$ My take is that no one knows enough about what the physical implementation of a qubit to make any such broad statement; the paper may estimate what it looks like what their technology would yield; there are several other competing technologies that may ultimately win out (or possibly something else that hasn't been invented yet). In any case, if we're talking about (say) the NSA, a dedicated power plant to run their quantum computer would certainly not be an obstacle... $\endgroup$ – poncho Nov 7 '17 at 17:56
  • $\begingroup$ As far as I know there are no bounds on energy consumption per qubit (would love to know if there are any). Classical bits have a "power consumption" equivalent to the necessary error correction rate times the Landauer limit if one maintains them, and of course zero if one allows thermodynamics to slowly randomize them. I expect something similar for qubits. $\endgroup$ – Anders Sandberg Nov 7 '17 at 22:08
  • 1
    $\begingroup$ Where did you hear this strange claim? $\endgroup$ – Nathaniel Nov 8 '17 at 8:37
  • $\begingroup$ @Nathaniel It's one of these things I've heard enough times, I don't remember the exact source. Probably something I heard at PQCrypto2017 or PQCrypto2016 conferences. $\endgroup$ – Mike Ounsworth Nov 8 '17 at 12:09
  • $\begingroup$ Hmm, well it seems a very strange claim to me. There are a number of proposed ways to implement cubits, but none of them would take any significant amount of power at all. In fact, on a fundamental level they can't take power, because a cubit has to be strongly isolated from the outside world to maintain coherence. So the power would only be for external things like cooling, and while that may well be significant, it's hard to imagine it needing a dedicated power plant. I suspect nobody will be able to say much more than that without knowing more about why the claim was made. $\endgroup$ – Nathaniel Nov 8 '17 at 12:17

Qubits will need to be error corrected. The classical compute needed to do this is a dominant cost in power.

Suppose you need a million physical qubits, and 1 cpu can manage the error correction for a hundred of them with sufficiently low latency. Then you need a megawatt of power to run the ten thousand CPUs keeping the quantum computation on track. That's 1W per qubit, plus or minus a factor of 10 since all the numbers are rough.

Obviously that would get better as people worked on the problem and made specialized hardware.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.