Imagine you're shipping packages through a delivery network. You wonder: if I tie packages together in clever bundles across multiple shipments, can I squeeze more total cargo through the same trucks? For abstract theoretical trucks, Hastings showed in 2009 that yes, sometimes bundling helps — but only for absurdly large, physically implausible trucks. This paper asks: for the trucks that actually exist in nature, does bundling ever help? The answer is no. The committed claim: for all phase-covariant quantum channels — the broad family covering amplitude damping, generalized dephasing, and damping-dephasing channels — entangling codewords across independent channel uses cannot increase the classical capacity. This is the additivity of the Holevo information for these channels. The paper introduces a new technical tool, the Schur-weighted block entropy inequality, to prove it. To understand the ladder: the minimum output entropy conjecture (additivity for all channels) was disproved by Hastings in 2009, but only for channels requiring enormous Hilbert space dimension — too large to correspond to physically relevant systems. Subsequent work by Ding-Junge-Li and Siddhu-Griffiths established additivity for all qubit channels and generalized amplitude damping channels specifically. This paper subsumes and extends both results to the entire phase-covariant family, which is the largest physically motivated class of channels treated so far. The architecture is pure mathematical proof, not simulation or experiment. The key structural innovation is the Schur-weighted block entropy inequality — a representation-theoretic tool that decomposes the output state of entangled codewords using the Schur-Weyl duality between the symmetric group and the unitary group. The method exploits the covariance structure of phase-covariant channels: because these channels commute with diagonal unitaries, the output entropy computation factors cleanly through irreducible representations. This is fundamentally an algebraic argument, not a computational one. Integrity here is unusually strong for quantum information theory. The entire proof has been mechanically verified in Lean, with code available on GitHub. This is not 'code released' in the sense of a software artifact — it is a machine-checked mathematical proof. The companion repository lets any reader with a Lean installation verify every step. For a proof paper in mathematical physics, this is close to the gold standard of validation. The milestone question is about scope: phase-covariant channels are physically important but do not exhaust all low-dimensional channels. The natural next target is all qudit channels (dimension d ≥ 3), or — more ambitiously — identifying the precise dimensional threshold at which Hastings-type counterexamples become possible. The current work handles all channels that commute with diagonal unitaries; extending to channels with less symmetry is the frontier. The obvious experiment not run: extending the Schur-weighted inequality to channels without covariance structure. The honest read is (a) — the technique is specifically tailored to exploit phase covariance, and generalizing it to non-covariant channels likely requires fundamentally new ideas, not just more effort. The authors are transparent about this: the method's power comes precisely from the symmetry it can exploit.
