0
I’ve seen several people that worked on quantum computers for years say that even when it works, we have very little software to run on it.
The software gap is real and, by Google Quantum AI's own admission, now the bigger bottleneck — finding problems where quantum beats classical and turning them into working applications is harder than building the hardware.
Why it matters: Billions are flowing to hardware companies like those in this story, but payoff requires software that doesn't yet exist at scale.
- Google Quantum AI researchers explicitly named the application layer — not qubits — as the field's most underfunded and underestimated challenge, citing two gaps: identifying problems with genuine quantum advantage, and mapping those to real-world use cases.
- IQM's State of Quantum 2025 report concluded the industry must solve software platforms and talent shortages, not just qubit counts, to scale commercially.
- Quantum software must currently be rewritten for each hardware architecture (superconducting, trapped-ion, photonic), so there is no shared software stack — the fragmentation that plagued early computing before UNIX and x86 standardized things.
- The closest thing to a near-term 'killer app' is quantum chemistry simulation (drug discovery, materials science), but as of mid-2026 no quantum solution is commercially indispensable in any sector.
Room for disagreement
- Hardware-first investors argue the software gap is expected and historically normal — PC software didn't exist before PCs did, and the stack will follow once machines are reliable enough to program against; the counter-view is that classical computing had a clear general-purpose instruction set from the start, whereas no equivalent abstraction layer exists yet for quantum.
AI-generated with web search — may be wrong. Check the linked sources.
