We present a comprehensive analysis of the etymological origins of all first names given at birth in France between 1900 and 2024, using the complete INSEE civil registry dataset (87 million births, 48,516 unique names). Each name was classified into one of 20 etymological origin categories using a large language model (Claude Haiku 4.5) operating as an automated onomastic classifier. Our analysis reveals four major structural shifts: (1) a sustained decline of names with Germanic etymological roots, from 28% of births in 1920 to 8% in 2024; (2) a collapse and partial recovery of Hebrew/Biblical names, peaking at 40% in 1946 before stabilizing at 23%; (3) a steady, quasi-linear rise of names with Arabic etymological origins from near-zero in 1950 to 16% in 2024; and (4) a monotonic increase in the Shannon diversity index of name origins across the full period. Monte Carlo projections (10,000 trajectories calibrated on 1990–2024 volatility) produce 90% intervals for 2050. The full classification dataset (48,516 name–origin mappings) and analysis code are included as supplementary materials. A shorter French-language version is included as an additional file. An interactive visualization of these results is available at https://yukicapital.com/french-first-names-origins