We present parameter-free predictions of the nuclear modification factor, R π AA (pT , s), of high pT pions produced in Pb+Pb collisions at √ s NN = 2.76 and 5.5 ATeV based on the WHDG/DGLV (radiative+elastic+geometric fluctuation) jet energy loss model. The initial quark gluon plasma (QGP) density at LHC is constrained from a rigorous statistical analysis of PHENIX/RHIC π 0 quenching data at √ s NN = 0.2 ATeV and the charged particle multiplicity at ALICE/LHC at 2.76 ATeV. Our perturbative QCD tomographic theory predicts significant differences between jet quenching at RHIC and LHC energies, which are qualitatively consistent with the pT -dependence and normalization-within the large systematic uncertainty-of the first charged hadron nuclear modification factor, R ch AA , data measured by ALICE. However, our constrained prediction of the central to peripheral pion modification, R π cp (pT ), for which large systematic uncertainties associated with unmeasured p+p reference data cancel, is found to be over-quenched relative to the charged hadron ALICE R ch cp data in the range 5 < pT < 20 GeV/c. The discrepancy challenges the two most basic jet tomographic assumptions: (1) that the energy loss scales linearly with the initial local comoving QGP density, ρ0, and (2) that ρ0 ∝ dN ch (s, C)/dy is proportional to the observed global charged particle multiplicity per unit rapidity as a function of √ s and centrality class, C. Future LHC identified (h = π, K, p) hadron R h AA data (together with precise p+p, p+Pb, and Z boson and direct photon Pb+Pb control data) are needed to assess if the QGP produced at LHC is indeed less opaque to jets than predicted by constrained extrapolations from RHIC. PACS numbers: 12.38.Mh, 24.85.+p,