Quantum chemistry simulation -- the most-cited 'killer app' -- may actually have very limited quantum speedup

For over a decade, quantum chemistry simulation (calculating molecular ground states, reaction dynamics, and material properties) has been promoted as the primary 'killer application' that would justify quantum computing's enormous cost. But recent research published in 2025-2026 indicates that two of the most popular quantum algorithms for chemistry problems -- the variational quantum eigensolver (VQE) and quantum phase estimation (QPE) for molecular simulation -- may offer very limited practical advantage over improved classical algorithms, even on future fault-tolerant quantum hardware. Why it matters: Because the most-cited commercial justification for quantum computing is weakening, pharmaceutical and chemical companies that allocated R&D budgets for quantum chemistry exploration (Merck, Roche, BASF, and others) are pulling back or pausing programs, so quantum hardware companies lose their most compelling sales narrative for enterprise customers, so the gap between quantum computing investment and quantum computing revenue widens further, so public market quantum computing companies (IonQ, Rigetti, D-Wave) face increasing pressure from investors who see no path to product-market fit, so the entire quantum computing ecosystem risks a credibility crisis similar to the AI winters of the 1970s and 1990s. The structural root cause is that classical algorithm researchers have not been standing still while quantum computing has developed. Tensor network methods, density matrix renormalization group (DMRG), and machine-learning-enhanced classical simulation have improved dramatically, closing the gap that quantum algorithms were supposed to exploit. The quantum advantage window for chemistry was estimated based on 2015-era classical algorithms, and those estimates were never systematically updated as classical methods improved.