This study examines how patent portfolios relate to market capitalization in an economy where intangible assets account for roughly 90% of corporate value. Using a 10-year panel (2015-2024) of 40 public firms across technology, pharmaceutical, and consumer sectors, I estimate OLS, firm fixed-effects, and generalized estimating equation models within the Hall-Jaffe-Trajtenberg valuation framework to test whether citation-weighted patent quality outperforms raw patent quantity as a predictor of market value. Results show that citation elasticity (β₃ = 0.26-0.31, p < 0.01) is approximately 2-3× larger than the elasticity of patent productivity (β₂ = 0.06-0.12), indicating that patent quality is substantially more informative for valuation than simple counts. Technical founder-CEOs sustain about 22% higher R&D intensity and exhibit a positive interaction with citation quality, while geographic filing breadth serves as a cost-effective quality proxy: only 3.3% of Chinese patents are filed internationally versus more than 70% for advanced economies. Sector heterogeneity is pronounced, with the patent-value relationship strongest in pharmaceuticals (ρ ≈ 0.71), moderate in semiconductors, and weak and often statistically insignificant in consumer goods. Case studies of NVIDIA's decade-long CUDA investment, Intel's forgone AI platform bets, and a projected 2025-2030 pharmaceutical "patent cliff" illustrate the mechanisms by which technical leadership and AI-enabled drug discovery reshape the economic returns to patents. The findings imply that conventional patent metrics systematically mismeasure innovation value and suggest that investors, corporate strategists, and policy makers should overweight citation velocity and geographic breadth-rather than raw counts-when evaluating innovation-driven market value.