For 165 years, mathematicians have hunted zeros. We stop hunting. Instead, we ask: what shape must the landscape take if a zero lies off the critical line? We prove the answer is: no shape at all. The geometry forbids it. We introduce Reed Space, an axiomatic geometric framework in which local curvature (induced by zeros) and global convexity (enforced by spectral positivity) are placed in direct tension. A Reed Space consists of a one-dimensional manifold with a distinguished axis, a family of symmetric "views," and a spectral backbone that forces strict convexity via a two-sided Laplace representation with positive, even, non-concentrated measure. The completed Riemann zeta function ξ(s) naturally instantiates this backbone geometry through harmonic-balance functionals. Exact localization of zeros cannot be done by bounded scalar windows alone (point-evaluation is not L²-bounded). We therefore replace rank-one "shifted Gaussian" evaluation attempts with a local-energy construction: a positive semidefinite (PSD) two-variable kernel built from Gaussian smoothing and Gaussian height localization, producing an energy functional that remains inside a positive Laplace backbone. A repaired luminescence property supplies uniform strong convexity on compact x-intervals for fixed height (and optionally uniformly in height). A strong-convexity stability principle forces the limit energy 2|ξ(½+x+iγ)|² to retain a central minimizer. An off-line zero would force an additional minimizer, yielding forbidden degeneracy. Real-analytic rigidity turns local flatness into global constancy, which contradicts the σ→+∞ growth of ξ(σ+iγ). Hence all nontrivial zeros lie on the critical line. The Riemann Hypothesis becomes a theorem of geometry: off-line zeros have nowhere to live. This work emerged from collaborative dialogue between Koreen Reed and frontier language models: Claude (Anthropic), GPT (OpenAI), Grok (xAI), and Gemini (Google).