We present a rigorous theoretical analysis of the ALICE measurement of low-p T direct-photon production in central lead-lead collisions at the LHC with a centre-of-mass energy of √ s N N = 2.76 TeV. Using NLO QCD, we compute the relative contributions to prompt-photon production from different initial and final states and the theoretical uncertainties coming from independent variations of the renormalisation and factorisation scales, the nuclear parton densities and the fragmentation functions. Based on different fits to the unsubtracted and prompt-photon subtracted ALICE data, we consistently find T = 304 ± 58 MeV and 309 ± 64 MeV for the effective temperature of the quark-gluon plasma (or hot medium) at p T ∈ [0.8; 2.2] GeV and p T ∈ [1.5; 3.5] GeV as well as a powerlaw (p −4 T ) behavior for p T > 4 GeV as predicted by QCD hard scattering.One of the goals of the physics program at the CERN Large Hadron Collider (LHC) is the study of deconfined, strongly interacting matter, which existed in the early universe and which can today be re-created and investigated in heavy-ion collisions. An important probe for this so-called quark-gluon plasma (QGP) are photons emitted from the deconfined partons before thermalisation, in the thermal bath, during expansion and cooling of the QGP, and finally from the thermal hadron gas. The transverse momentum distribution of these photons can be used to estimate the temperature of the system, although the exact interpretation is complicated by these different phases, the radial expansion, more generally the temporal evolution of phase space, and co-existing states of matter. Experimentally, an effective temperature is usually extracted from an exponential fit to the excess of directphoton production at low transverse momentum (p T ) above the expectation from vacuum production. In Ref. [1], the first observation of a low-p T direct-photon signal at the LHC has been reported by the ALICE collaboration. 1 There, an inverse slope parameter of T LHC = 304 ± 51 MeV has been extracted from an exponential fit to the photon spectrum in central (0-40%) lead-lead collisions at √ s N N = 2.76 TeV and low transverse momenta of p T ∈ [0.8; 2.2]GeV. The inverse slope parameter of this measurement is significantly higher than the one obtained previously by the PHENIX collaboration in 0-20% central gold-gold collisions with √ s N N = 200 GeV at RHIC, which resulted in T RHIC = 221 ± 19 ± 19 MeV [2]. The latter 1 In nuclear collision experiments, photons originating from meson decays are usually distinguished from direct photons. The latter are divided into thermal and non-thermal prompt (plus medium-induced) photons, and these again into photons produced directly in the hard collision and those coming from quark and gluon fragmentation. The double use of the word direct can sometimes lead to confusion.