One of the things that makes it so tremendously hard to be a graduate level physics student in fundamental modern physics is that there are a vast number of proposed theories, hypothetical particles, and so on, out there, each of which tends to require some absurdly esoteric kind of mathematical understanding. You must learn not just general relativity and the Standard Model, but also their inconsistencies and all of the experiments that rule out or tend to establish deviations from them.
Someday, and I am optimistic that it will come by the time that my grandchildren are physics students advanced enough to learn quantum mechanics and general relativity, we are going to have all of this worked out, we will have a theory of everything, or at least, a few consistent theories that explain everything at the fundamental laws of physics level, even if they are somewhat ugly and not unified in the grand unified theory or theory of everything sense. The puzzles that remain are few, and the progress we have made so far is great. Dark matter is something of a puzzle, but we can describe limitations on its effects quite well from experiment. Dark energy has a perfectly good answer right now (the cosmological constant of general relativity) that we just need to confirm is the only possible correct answer. There may be only one more (or fewer) fundamental particles less to discover. As almost all of the parameters of the Standard Model are finally pencilled in from experimental data, we are finally starting to clue into the connections between the fermion mass matrix, the CKM matrix, the PMNS matrix, and the coupling constants of the three Standard Model forces, parameterizing them much more efficient as we learn their relationships. Unifying general relativity and the Standard Model is rather less of a chore than determining that the Standard Model is complete and accurate and forms some sort of coherent whole, since honestly, there are very few occasions when both have to be brought to bear and the stylized situations where they do can be solved as special cases with ad hoc intuition guided by experiment if need be. It will make more sense, as well, because the lingering questions will be addressed and will fit better into an overall whole.
When this happens, a vast amount of the stuff that a graduate level physics student needs to know these days will become irrelevant. We can't be quite sure which pieces in particular will become irrelevant, but we can be pretty sure that our hapless graduate student will not need to master both compacted ten dimensional space and the discrete mathematics that goes into loop quantum gravity; both knot theory and Klein-Kaluza topologies; both several MOND theories and several dark matter theories. Our hapless graduate student may need to learn one of each, but probably not both.
Similarly, our hapless graduate student will not need to learn the Standard Model, two or three variations on supersymmetry, technicolor, the SM4, and half a dozen other extensions of the Standard Model. Learning the new and improved Standard Model and that math that goes with it once will be good enough.
Likewise, it is extremely likely that any resolution of the current conundrums of the Standard Model will greatly reduce the nineteen or so parameters that the current model involves, with values that change every other physics conference and take a handbook that runs hundreds of pages to determine. There will be highly precise values for half a dozen constants (a bit like the number of the classical Newtonian/Maxwellian model), and everything else will be calculated exactly from them. The necessary fundamental constants will fit on the book flap of the textbook (except that our hapless graduate student probably won't be using a physical paper textbook anymore). All the little boxes where particles and couple constants and equations go will be filled and there will be no room in the margins for meddling and revisions. The detailed properties of every possible composite particle, calculated out to several significant digits from the handful of fundamental constants, with all their excitations, will fit in a few dozens pages of the CRC Handbook of Chemistry and Physics, along with the periodic table, the table of isotypes and long lists of organic and inorganic chemicals with their properties.
Honestly, I think that it is all going to look surprisingly familiar.
I think it is quite a bit more likely that SUSY and a fortiori, String Theory/M-Theory, will be disproven than it is that it will be confirmed in any form. There may or may not be a Higgs boson, and if there is one, there is a good chance that there will be just one Standard Model one (as the Standard Model seems to manage just fine with a single vev value for the Higgs field). There may be a fourth or higher generation fermion, but there is a good chance that there isn't. I expect no new fundamental bosons besides the Higgs. There might be right handed neutrinos, but there is a good chance that there aren't. A suspect that the trick to getting the coupling constants of the Standard Model forces to align is to tweak the equations (or the way that we approximate them numerically) slightly rather than to doube the number of particles in the zoo. Whatever we find to solve the theoretical problem that the as yet undiscovered Higgs boson has presented us with, I very much doubt that the solution will involve technicolor, even walking technicolor. I very much doubt that the theory that we end up with will include more than five dimensions or that it will include any compactified dimensions or branes. There might, or might not, be a graviton in the final theory; I think it is more likely that quantum gravity, if necessary, will arise from the non-point-like architecture of space-time. There might be one layer of particles more fundamental than the ones we know now.
I wouldn't be surprised to see a reformulatization of QCD that is equivalent to or nearly equivalent to the current theory that makes it easier to calculate. I wouldn't be surprised to see a couple of new and profound new theories that address the issues of the non-localities found in modern physics, the arrow of time, and better define the role of the observer in the outcome of physical processes.
I wouldn't be surprised to see dark matter resolved with a newly discovered "warm dark matter" candidate like a sterile neutrino or glueball, nor would I be surprised to see it resolved with refined calculations using the existing theory of general relativity or a somewhat modified theory of general relativity, perhaps flowing naturally from a quantum level adaptation of it.
Footnote: The vacuum expectation value of the Higgs field is 246 GeV. The Higgs boson, if it exists, is more than 114 GeV and probably less than 130ish GeV. Just to be cute, because it is possible, I will make a prediction. If there is a Higgs boson, it will have a mass of exactly one half of the vev of the Higgs field, i.e. 123 GeV/c^2. If I'm right, you heard it here first. If not, my batting average is still at least as good as the ordinary theoretical physicist.
Of course, if the LHC does discover a single Standard Model Higgs of spin zero, 123 GeV, and no electrical charge or color charge, this would go a long way towards putting all of those theoretical physicists out of business be eliminating much of the motivation for them to come up with new theories. The end of LHC's run may not strictly speaking rule out Sting Theory or SUSY or Technicolor, just as we haven't absolutely ruled out magnetic monopoles and proton decay and flavor changing neutral currents to absolute mathematical certainty.
But if the LHC does nothing but discover the Higgs and confirms the Standard Model until its run is complete, without finding another wiff of "new physics," they aren't going to build another bigger and better particle accelerator to replace it any time soon. If an apparently theoretical sound way to set the values of the PMNS matrix and neutrino masses from LHC data about quarks that if confirmed by all experiments to data, even neutrino experiments are going to start to look a little pointless, like paying people to reconfirm the universality of the force of gravity or F=ma at low masses and speeds.
If LHC runs out of "new physics" and finds a single Standard Model Higgs boson, the only games in town are going to be figuring out dark matter, coming up with a way to make sense of cosmological inflation, coming up with ways to reconcile gravity to the rest of the Standard Model, and improving our ability to do QCD and GR equations in less stylized contexts.
There is an urgency in fundamental physics right now because the foundations of physical science are embarassingly broken. But, the current cracks in the edifice, while troubling, are largely cosmetic. None of the other important unsolved problems in science at the moment depend upon breakthoughs by people engaged in fundamental physics. We know enough about how the world works to manage, brute force style with an ugly theory that works.