We have successfully put 37 trillion nanobots the size of a red blood cell into
the experimental group test subjects' willing participants' bloodstream, but without a power source, the nanobots clot up the bloodstream and don't carry oxygen cause mild side-effects that last more than 4 hours. EnerThrive© Inc. is committed to avoiding lawsuits providing our customers with the best possible experience, so we've reached out to the uneducated scum public for free intellectual property ideas of how to power the nanobots.
What would be a feasible way to power trillions of nanobots? These nanobots would live in the bloodstream, so they never should run out of power. Powering the bots includes both acquiring the power, and storing the power. Assume that the nanobots mostly use technology available today (except scaled down, of course), though there can be some minor-moderate advances in a few areas. The nanobots (and their power) should last at least a whole lifetime. Whatever power source also must be very cheap (because you need 37 trillion of it). A handwavium-boson-tachyon nanofusion engine should be a last resort. Also, no magic.