Jay,
Why is that..? What did you read about nanotech that made you think this? Anyway, your body is made of nanotech and it isn't very energy intensive.
Rearranging many trillions of covalent bonds per second would be very energy intensive and would also release a lot of heat that would require cooling. Enormous computational power is also required to run the nanofactory.
Yes, very energy efficient devices can be built with nanotechnology and we could save a great deal of energy by making the investment upfront (e.g. nanomanufacturing ultra efficient light bulbs and engines) but what we can build or recycle will ultimately be limited by how much energy we have.
In light of this my view is that nanotechnology will not solve our energy crisis. Breaking and rearranging bonds will never get us more energy than we put in.
Once we have robust nanotech humans will consume more energy not less.
We must look to the solar system to get the energy required without trashing our environment and spending earth resources that should be reserved for future generations.
Biological nanosystems are low energy intensity, but are not very robust for that reason. High energy particles and molecules can seriously damage them rather easily (like how cooking precipitates proteins out of solution) for the simple reason that in order to operate at a low energy level, their molecules need to be rather loosely bound. Extremophilic bacteria must replace elements in their proteins, like iron, with elements like tungsten in order to more tightly bind protein molecules together, but doing so prevents those proteins from being able to function at room temperature. Wet nanotech will not be heat intenstive, but will be limited in computational power and repair capabilities to within a reasonable range of organically evolved natural systems.
If chemical bond computation is not going to be used (as biological systems do), then either mechanical, electrical, or optical systems must be utilized. mechanical systems generate lots of heat by the friction of their moving parts, and may even generate some heat from the structural stress of resisting various forms of casimir forces within their structures. Electrical obviously generates resistance related heat, and optical, which generates heat by exhausting the quanta of photons in computation, and emitting the waste as black body radiation, may be the most efficient as well as the most robust.
Some nanosystems may use a mixture of these processes, but their robustness will be limited to the weakest of their links.
Nanotech will solve the energy crisis through efficiency. Integrated silicon circuitry is incredibly inefficient in terms of flops per watt-hour. This is a pie cutting exercise: if you make your energy consumption infrastructure 50% more efficient, you double the lifetime or capacity of your energy resource base.
As an example, as some know, I invented a product to retrofit exit signs about 15 years ago, using EL lamps. 97% energy efficient, we were able to retrofit 40 watt incandescent exit signs with 0.31 watt EL lamp retrofit kits, providing essentially the same illumination, and often with improved quality due to the even emission of light over the surface of the lamp, and its emission within a narrow band of the visible spectrum most efficient at being observed by human eyes through smoke. This product has saved many millions of kilowatt hours of energy. You think, 40 watts isn't so big, but those exit signs run 24/7, while most normal lights only operate during working hours. i.e. a 40 watt exit sign consumes as many watt hours as a 80-120 watt fluorescent light fixture. There are hundreds of millions of exit signs in the US alone.
The whole reason for the low oil prices of the 1990's was the vehicular efficiencies developed in the 1980's. The current push to transition to hybrid vehicles will lead to a similar reduction in oil prices and expansion of the practical capacity of the known oil reserves. Going from a world of 25 mpg cars to 50 mpg cars means rather than 75 years of oil at 80 billion barrels a year consumed, we have 150 years of oil at 40 billion barrels a year consumed (and future efficiency gains mean further doublings).
There is no energy "crisis". We are in a point of time in which we are in a temporary transportation efficiency singularity, and will shortly transition to a new technology (hybrids) to begin a new asymptotic curve, as Kurzweil has explained.