PUH Theorem 368 — Every Planck Core Makes Spacetime at Its Poles, but All Their Shells Together Supply 10⁻¹⁹ of the Late Push and Run on the Wrong Clock: the Rebound Drives, Gravity Brakes

Photonic Universe Hypothesis (PUH) — Calculation and Record. THE PICTURE THIS NOTE RECORDS (the author, 24–25 September 2026). Planck cores are NOT black holes and contain NO singularities: each is a bubble of a different phase (T359) whose interior has no lattice, no nodes and no propagation (T180), bounded by a Shell at most 7.7 per cent outside its gravitational radius today (T311, T300, T315); where continuum physics would diverge, the lattice crystallises a core instead (T173); cores never evaporate or shrink (T206, T336, T361). Astronomy catalogues them as black holes. EVERY PLANCK CORE MAKES NEW SPACETIME AT ITS POLES — a smaller version of what the final core did at the rebound (Three Boundaries; T208; T242). The rebound drives the expansion: two polar jets of high-energy photons, matter from one pole and antimatter from the other, become two universes with a throat of the Super Polariton Field (SPF) between them; the SPF presses on our cone and keeps it conical; lattice tension rises and slows the growth (Regime I). Cores drain spacetime at a limited rate but do not push or pull the edge. RESULT 368.1 (how much). The rebound released a universe the final core had stored; an ordinary core cannot release what it holds, so its poles can only pass on what its Shell processes, and throughput is rate times area (T213) with the rate capped at one Planck layer per Planck time, a speed c (T210). The black-hole census (entropy 3.1×10¹⁰⁴ k_B, Egan & Lineweaver 2010) gives 3.24×10³⁵ m² of horizon, so 3.76×10³⁵ m² of Shell. AT THE CEILING ALL PLANCK CORES TOGETHER MAKE 1.13×10⁴⁴ m³/s AGAINST THE 1.00×10⁶³ m³/s THE LATE PUSH NEEDS: 1.1×10⁻¹⁹. The Shells would need 8.9×10¹⁸ times as much area — 1.4 times the area of the observable universe's own boundary. The drain is capped by the same number (4.9×10⁻²⁰ of the volume growth). Locally the poles out-create the Hubble flow only within 0.096 pc of Sgr A* and 12.6 pc of M87*. RESULT 368.2 (when). A steady push needs output that grows with the volume of the universe, (1+z)⁻³. The cores' Shell area barely grows and their activity peaked ten billion years ago (Madau & Dickinson 2014 shape). If pole creation supplied today's push, the push would have been 2.3 to 9.9 times stronger five billion years ago; the CMB then forces H₀ to 39.4–55.5 and DESI DR2's rulers give χ² = 175 to 7,709 against 31.0 for a constant push. THIS HOLDS WHATEVER EACH CORE'S EFFICIENCY, SO IT HOLDS FOR CREATION ALONG THE JETS TOO. RESULT 368.3 (who pays). If new space carries the push's measured energy density, filling the space of the last five billion years took 0.686 ρ_Λc² per cubic metre of today's space; every baryon converted completely would pay 10.5 per cent, and everything the cores ever released as light and jets 3×10⁻⁷ to 1.4×10⁻⁶. Jets cannot pay; the energy is the lattice's own (T366: not yet supplied by PUH's field equations), or the SPF's (T356, a proposal), or new space carries none. RESULT 368.4 (what drives the expansion). T231: the rebound's outward kinetic energy drives the conical expansion. With pressure weighing (T312) and radiation losing energy by P dV work on the lattice (T364), the braking law has an exact first integral (residual 1.3×10⁻¹⁵), so an escape-speed launch reproduces the measured expansion rate to 1 per cent from one second until z = 3.7, 12.1 billion years ago: nucleosynthesis, the CMB (sound horizon 144.26 Mpc in both) and structure growth pass, on three conditions — pressure weighs (otherwise 29 per cent slow at one second, 20σ), the launch matched escape speed to 1 part in 10⁶², and the cascade was over before one second (≤ 3.8 per cent extra radiation). RESULT 368.5 (the late push is required). DESI DR2's rulers alone exclude braking-only at 3.6σ (χ² 20.6 for 4 d.o.f.; with a push 4.0); with today's rate 9 to 11σ; the speed-up began near z = 0.75, about 6.8 billion years ago. The supernova age-bias debate (Son et al. 2025; Wiseman et al. 2026; Chung et al. 2026) does not remove it: the result uses no supernovae. RESULT 368.6 (the launch). No measurable spin (below 3.3×10⁻³⁷ of the expansion rate at the rebound; T136's quantum passes by about 112 orders), growth the same in every direction (the SPF's shape-keeping pressure), at most 4.4 per cent of matter added since the CMB, at most 10⁻⁶ of the final core's magnetic-field energy in space, antimatter in the other universe. Escape-speed precision, η and the primordial ripples are required and not yet derived. RESULT 368.7 (where the push can be made). Shape alone and edge sites give coasting; sites that multiply with the volume give a Λ-like push. T362's scarce-mass trigger: a push following below-average-density regions beats a constant push with no extra parameter (χ² 23.7 against 31.0); a near-emptiness trigger switches on too late (171, 944). Voids plus rising tension reproduce DESI's rise-then-fade (10.37 against 10.19) only with tension today at a third of tearing, which T300's floor (X ≥ 14 today, T315) forbids; within the floor the best is 16.7. A TESTABLE FORK: EHT Shell imaging and failed-supernova counts with redshift (T351's threshold 3.14 M☉ today on the floor branch; 2.26 M☉ today and 3.34 at z = 0.5 on the other). RESULT 368.8 (the CMB). Relic light of the early plasma, as Regime I requires: edge light is excluded by its energy (stretch 1.7×10³² from 9E_P), its shape (an order-unity lopsided sky), and its history (5.08 ± 0.10 K at z = 0.89 against 5.15 predicted). The hot phase began 13.79 billion years ago; the oldest globular clusters fit inside it. "Far older" can only mean earlier cycles. CORRECTIONS. T178's drain-as-engine reading (and T240 §5, T252, T296 as expansion laws) WITHDRAWN by the author; T179's core count dead as written (7×10⁶¹ times all the matter in view); T165's cone law describes the edge only (5.2×10⁷ too slow at nucleosynthesis); T159/T232 refractive redshift dead as written (118 orders); T136 arithmetic slip (5.4×10⁻¹⁴¹ rad/s, passes); T216 narrowed; the Z3 phase paper's step to η invalid; T358's flow resolution rests on T178's lock, which is open. KILL-CONDITIONS: a Shell passing lattice faster than the Planck rate, or a core releasing stored lattice without shrinking (368.1 only); a Planck-core population absent from the black-hole census with 1.4 times the area of the universe's boundary; a push found to track galaxy and core activity; pressure not weighing (engine fails BBN at 20σ); DESI's fade confirmed with T300's floor holding (site question reopens). NOT CLAIMED: that pole creation does not happen — THE ARCHIVE'S CLAIM STANDS, AND THIS NOTE GIVES ITS SIZE AND ITS TIMING; that Planck cores are shown to exist or not to be black holes; that the void push is established; that the voids-plus-tension fit is a success; that the launch precision, η or the ripples are explained; that the universe is no older than its hot phase.