Less than than 1eV/cm[sup]3[/sup], or in perhaps more familiar units, less than 10[sup]-13[/sup] Joules per cubic meter, on average in the universe. (The energy density locally can be a few orders of magnitude greater when there happens to be a strong wave passing by, like with the first LIGO detection.)
Doesn't sound like much? Indeed! This is only a few parts in 100,000 of the critical density of the universe (the critical density being about 5 proton masses or 10[sup]-9[/sup] Joules per cubic meter). For comparison, matter makes up about 30% of the critical density, and dark energy makes up about 70%.
How do we know? Gravitational waves contribute a form of energy density to the universe that we group under "radiation density". That also includes electromagnetic radiation, neutrinos, and any particles moving at highly relativistic speeds. We group these together because they dilute with expansion in the same way -- their densities all decrease according to the 4th power of the size of the universe, whereas matter density decreases with the 3rd power of universe size.
The individual densities of matter, radiation, and dark energy are all measurable from the Cosmic Microwave Background radiation. Specifically, by the "CMB angular power spectrum", which is explained in more detail in the two PBS Space Time videos linked here. The total density of matter, energy and radiation together also affect the curvature of the universe. The total density is about 1.02 times the critical density (with an error of about .02), of which radiation is a few parts in 10[sup]5[/sup] -- a tiny amount.
Most of the radiation density comes from the CMB photons: about 0.25 electron volts per cubic centimeter. The cosmic neutrino background also contributes some energy density. So there is very little room left for gravitational wave energy in the universe. This probably sounds surprising given all the reports of how much energy is carried in the waves detected by LIGO, and what I just said about supermassive black hole mergers radiating hundreds of thousands of solar masses worth of these waves. But such strong gravitational wave sources are rare events, whereas the CMB and CNB uniformly fill every cubic meter of space.
Another possible cause for confusion is that the energy carried in gravitational waves is usually compared to the energy of starlight, and specifically in terms of the peak power of the signal (which lasts less than a second) vs. total emission of starlight in the same amount of time. But even all the combined starlight (over all time) in the universe is a tiny fraction of the energy of the CMB radiation, which again is only a tiny fraction of the total mass/energy density of the universe.
Another thing we can do is calculate what we expect the energy density of gravitational waves to be, based on our understanding of the sources. There is a lot of research in that area. Here's an example. We also had some fascinating discussion of the contribution from exoplanets earlier.
Yes, but not very well. Gravitational waves cause spacetime to stretch and squeeze with their passage, and that energy can transfer over to vibrations in matter, or a wavelength shift of light waves in the case of the lasers in LIGO. However, the "coupling" between the gravitational wave and matter is very weak. Most of the wave energy goes right through, just like neutrinos passing through a planet, or a 100m radio wave passing around/through your body.Is there any mechanism that can convert energy in the form of gravitational wave into another form of energy?
(Aside: the wavelength of the laser light in LIGO gets stretched or squeezed by the same amount that the distance between mirrors does, which might make us think it's impossible to measure the wave at all, but the trick is that the light enters and leaves the system much faster than the time it takes for the entire wave signal to pass by, because the gravitational waves LIGO measures have long wavelengths).
Great questions, by the way.