Deepblu has brought up three heterodox ideas, I thought I would mention my own responses.
First, what if there's no dark matter? I certainly think the success of MOND, a type of modified gravity, is very important; but it is not relativistic. There is a kind of superfluid dark matter theory, due to Khoury and Berezhiani, which can reproduce the features of MOND. One might also look for MOND to be a quantum gravity effect, perhaps involving extra degrees of freedom beyond the classical metric.
Second, what if there's no dark energy? This is more problematic in that QFT has vacuum energy and you expect vacuum energy to gravitate, so you actually expect there to be something like dark energy. Although then one has the problem that the expected magnitude of dark energy is vastly greater than what is actually observed. The anthropic answer is that there are positive and negative contributions to vacuum energy, and they happen to almost cancel because if they didn't, there wouldn't be galaxies, atoms, or life. More interesting is the idea that the QFT vacuum energy does cancel or almost cancel due to some symmetry, like a crypto supersymmetry. If it's only an almost cancellation, the observed dark energy can then be the residual vacuum energy. If the cancellation is exact, then the accelerating expansion has to come from somewhere else, such as a quintessence field.
Third is the heterodox idea that interests me the most precisely because I haven't thought about it: what if energy is conserved after all, in the true theory of quantum gravity? I am used to the usual line of thought, which is that you can get local conservation of energy in GR if you use pseudotensors. But how does the issue look in quantum gravity? This seems to be very little discussed. First of all, Noether's theorem already works a little differently in quantum field theory, compared to classical physics, because of the peculiarities of the quantum framework. And here is a perspective that is new for me: energy conservation is due to time translation, but the Hamiltonian of quantum gravity has no time evolution! How does that affect the attempt to reason about energy conservation? I have no idea.