arXiv2026
Classical general relativity predicts a singularity at the center of every black hole. We argue that it is never reached. We adopt the conjecture of Shaya (2026) that a metric element of non-Lorentzian signature is inadmissible: it cannot enter the Feynman sum over geometries, and so cannot belong to the physical manifold. Nothing beyond standard quantum mechanics is invoked. Signature is the only pointwise diffeomorphism-invariant content of the metric, so the criterion is local and coordinate independent. Transverse-traceless graviton vacuum fluctuations grow as $Δ_g = \sqrt{8/π}\,\ell_P/q$ on a scale $q$, and any finite-scale metric stress driving an eigenvalue through zero removes that element. When the surviving elements can no longer sustain connected, differentiable, causally propagating support, the manifold terminates on a quantum boundary $\mathcal{B}_Q$ at finite radius. For the vacuum Schwarzschild interior the tidal Weyl stress gives $r_{\rm qb} \simeq 1.0\times10^{-22}(M/M_\odot)^{1/3}$~m for the adopted curvature response $C_W=0.15$. For a spinning hole, the boundary forms much farther out: mass inflation at the inner horizon $r_{-}$ concentrates an exponentially growing interior mass into a thin layer whose thickness is set by the accretion rate and floored at one spatial quantum, and the counter-streaming null fluxes supply a trace-free Ricci stress along the radial axis alone. This caps the factor $n_{\rm qb}$ by which mass inflation can amplify the interior mass function, at between $10^{38}$ and $10^{69}$ for a $10\,M_\odot$ hole depending on its accretion history, and leaves the Cauchy horizon, the ring, and all deeper extensions outside the physical manifold. The Gibbons--Hawking--York term over this terminal slice yields a finite interior action, $S_{GHY}^{qb} \approx \frac{3}{2} n_{\rm qb} Mc^2\,Δt$.