I've seen some articles about biologically powering devices through a glucose fuel cell, but from what I can tell this would only provide power on the order of a few tenths of a watt.

While it's definitely an interesting scientific development, it probably wouldn't be ideal for powering something like a sixth-generation Intel i5/i7 NUC computer, which can draw between 38 and 77 watts under high load and approximately 17 watts when idling.

That being said, suppose there is a near-futuristic race of mammalians that have a computer of similar caliber implanted in their bodies. Consider that:

  • The power source shouldn't be implicitly fatal for the host to use - burning things or producing excessive amounts of ionizing radiation might be out of the question, for instance, but inadvertently decreasing blood glucose levels and triggering hypoglycemia due to starvation would be OK.
  • The power source would have to fit inside/against the body - something that's heavy/bulky enough to limit mobility wouldn't work.
  • The environment (insofar as available resources, atmosphere, etc.) is Earth-like.
  • The mammalians in question are intelligent enough to maintain the power source and computer themselves (to a reasonable extent - they're not Leonid Rogozov)
  • The computer itself is the only thing that needs power - there won't be any high-draw devices needing power as well, for example.
  • The amount of computational power required would eschew using low-power devices like the Raspberry Pi and the CPU type(s) available in most cell-phones.
  • The power source would not require direct interaction from the host while in use. (Pedaling a stationary bike would not be feasible, but drawing power through piezoelectric foot implants would.)
  • The power source would not necessarily need to be biological/evolutionary in its own right, so long that it is biologically compatible (i.e. titanium implant)

Would it be feasible to power these computers through biological, self-sustaining means? How might this power source affect the host?

Thoughts would be greatly appreciated.

  • $\begingroup$ Reading between the lines, you're looking to produce roughly 50W or so? Worth noting: that qualifies as "vigorous exercise" for human beings. $\endgroup$
    – Cort Ammon
    Nov 27, 2016 at 4:18
  • 2
    $\begingroup$ Maybe GMO sheep-monkey-dolphin-spider-electric eel-animal X" hybrid, the wool cause static charges to build up and is harvested as a power source. I'm brainstorming which animal part can double as capacitor... $\endgroup$
    – user6760
    Nov 27, 2016 at 4:20
  • $\begingroup$ [instinctively thinks of the Matrix] Harvesting the excess electrical potential from nerve impulses, so the mammal feels every sensation mildly less? $\endgroup$
    – Tony
    Nov 27, 2016 at 4:30
  • $\begingroup$ @CortAmmon 50W on a bicycle appears to be similar to a leisurely ride as opposed to vigorous exercise. Is it significantly different for on-foot exercise? $\endgroup$
    – Shard Wolf
    Nov 27, 2016 at 5:42
  • $\begingroup$ @ITWolf I took that number from a range on wikipedia: "Search Results A trained cyclist can produce about 400 watts of mechanical power for an hour or more, but adults of good average fitness average between 50 and 150 watts for an hour of vigorous exercise. A healthy well-fed laborer over the course of an 8-hour work shift can sustain an average output of about 75 watts." $\endgroup$
    – Cort Ammon
    Nov 27, 2016 at 5:59

4 Answers 4


I will investigate electrical energy generation in fishes. This is not the milisecond shocks that electric eels or rays produce, this is the constant electric field generated in most species of Mormyridae and Gymnotiformes. These are called the weakly electric fishes.

The electric fields produces by these fish are constant low level dipole-dipole fields that are used for detection of other creatures and communication. We are not interested in either of those appliations, just the power implications.

According to this paper, Apteronotus leptorhynchus (Brown ghost knifefish) generates a 'waveform' electric field that cycles between zero and 10 mV heat-tail dipole-dipole potential. This is effectively a form of AC power generation. Each pulse generates $1\times10^{-13}$ J of dipole energy with frequency of about 1000 Hz. The paper also notes that dipole energy increases as the square of the dipole charge (which is linearly related to potential drop).

Creating a similar continuous waveform field at 400 Hz for a related knifefish Eigenmannia virescens takes 2.5 mJ per second. This is only a 5g fish; the difference between the 2.5 mJ of energy spent and the ~0.0001 $\mu$J effective 'power' output is expended by the nervous system. Otherwise, the energy is directly converted from ATP at high efficiency.

Now an electric eel can generate 800V or more by simply taking 6000 or so of the same organ in the knifefish, and stacking them end-on-end to increase electric potential. So we can definitely increase potential output to whatever level we want. We probably don't need 120V Ac, but getting a medium voltage (24 volts maybe?) 60 Hz power source for out computer is totally feasible. It just depends on how much power we want.

So let make a bunch of leaps of faith (called assumptions, if I was actually doing real science). Lets say we can scale up small organs with little loss of efficiency, lets say we can stack them to get the voltage we need, lets say we can continuously operate them at 60 Hz. Lets say that nerve workings cost us 100% of the power output costs: 50 percent efficiency. Finally, lets say that since glucose to ATP is about a 40% efficient process, regular food to ATP is 25% energy efficient. Then lets say we can devote a proportion of our metabolic energy to electricity generation as a knifefish (~25 %).

In that case, we take a 2000 Calorie diet/metabolism, increase that by 25% (2500 calories) and use the difference at 0.25 * 0.5 = 12.5% efficiency to generate about 60 Calories of electric energy per day. That works out to $2.5\times10^{5}$ J per day or about 3 Watts.

What can you do with 3 Watts? Power a Rasberry Pi 2 baseline while under computing load. Not great, but if you make a few assumptions about increasing the power efficiency of computers, you could feasibly get to where you want to go.

So, eat 25% more, add a few thousand electro-organs, and you have your self a bio-computer.

  • $\begingroup$ Would the situation be any easier if the end-goal was generating DC instead of AC power? Most single-board computers (RPi and Intel NUC included) are themselves powered under DC (RPis use 5V over USB, x86-based SBCs generally use somewhere between 12V and 20V) with a separate transforming power supply. $\endgroup$
    – Shard Wolf
    Nov 27, 2016 at 5:53
  • $\begingroup$ @ITWolf I was strictly looking at bio-electric generation, and I didn't see any creature that generate a constant DC type electricity. I think you will have to accept AC and do a n AC-DC transform in your computing unit. But, like I concluded, I think you have enough power to do some serious computing. Going from 3W to 10W is just a few assumptions off and eating some more calories. $\endgroup$
    – kingledion
    Nov 27, 2016 at 12:51

I would look at using TEGs, thermoelectric generators. People produce a lot of waste heat anyway because of the need to maintain body temperature. Cover the body with TEGs and harvest some of that waste heat. The top and back of the head would be a good spot. It's energy people expend if captured or not and doesn't present any extra load on the body. Wikipedia link to Thermoelectric Generator

I was surprised to find this: Hand Heat Demonstrator and this Page with demonstration video

The generator is 40mm x 40mm and produces about 40mW. Assuming a useable body surface area of 1 square meter (1.4 for women 1.9 for men) Body Surface Area that comes out to (1/(0.04*0.04))*0.040 = 25 Watts.


keep in mind computers have shown a steady trend to use less power not more, by creating higher and higher efficiency. but really fats contain the same energy density as gasoline so it's not that unlikely. there is already research along those lines, specifically enzymatic biofuel cell here is a great review of current work for laymen ,https://www.electrochem.org/dl/interface/sum/sum07/su07_p28_31.pdf they even have a cell powering an ipod for a nice visual. One of their goal is small implants to power medical bionics. Microbial fuel cells work even better and can use a wider range of fuels but the bacteria would not make good implants, engineered human cells might work though.


The human brain draws about 12 to 16 W (depending on source). A human at rest dissipates between 65 and 100 W (depending on source, size and sex). An average resonably-fit human male can produce about 75 W of mechanical power for eight hours (Wikipedia), of course with an increase in metabolism and food intake. Thus it is not outlandish to assume that we could generate and dedicate, say, 15 to 20 W for use by our future embedded digital assistant, which should be plenty -- for example, a dual-core Intel Celeron N300 has TDP of 4 W). The conversion of the energy stored in the form of ATP into 3.3 V DC is left as a simple exercise to the interested bioengineers.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .