After fourteen months of research, engineering, and community-building, the DEDALUS project has officially closed. This final newsletter summarizes the key scientific results, publications, and community activities produced over the course of the project.
The headline finding: quantum advantage for data access, in the NISQ era, is more effectively realized through indexing rather than search.
Six numbers that tell the DEDALUS story
| 6x | Peak speed-up vs. PostgreSQL on individual queries |
| 86x | Faster lookups vs. Grover’s algorithm (QuaC indexes) |
| 50% | Of queries improved over PostgreSQL plans |
| 72% | Best-case QUBO variable reduction via pruning |
| 61% | QuaC accuracy on 7-qubit IBM hardware (vs. <1% for Grover) |
| 5+ | Peer-reviewed papers at top international venues |
Flagship systems
DEDALUS Framework | Hybrid Query Optimisation
The first fully automated end-to-end pipeline that takes a raw SQL query, formulates join ordering as a QUBO problem enriched with live database statistics, dispatches it to a classical or quantum backend, and executes the resulting plan against a live PostgreSQL instance. A connectivity-aware pruning strategy reduces QUBO dimensionality by 37.9% on average (up to 72%). On JOB, TPC-H, and synthetic workloads, the framework outperforms PostgreSQL on approximately 50% of queries at 100% plan correctness.
QuaC | Quantum Circuits as Indexes
QuaC introduces a new paradigm: treating a quantum circuit itself as a deterministic index structure. Rather than relying on Grover-style probabilistic search, QuaC compiles Knowledge Graph access patterns into reusable CNOT-based circuits whose lookups are deterministic, single-shot, and deployable on today’s NISQ hardware. On IBM quantum hardware, QuaC achieves 61% accuracy on 7 qubits where Grover collapses below 1%.
Quantum-PG-HIVE | Schema Discovery
Extends the PG-HIVE framework for hybrid incremental schema discovery in property graphs by replacing its Locality-Sensitive Hashing step with a QUBO formulation of balanced minimum cut. The result is cleaner, more interpretable, and more robust schemas — solved classically via simulated annealing today, ready for quantum annealing tomorrow.
What’s next – QC&DKM 2026
The project closes, but the community continues. The 2nd Workshop on Quantum Computing & Data/Knowledge Management (QC&DKM 2026) will be co-located with VLDB 2026 on September 4, 2026 in Boston, MA. Visit qcdkm.github.io/2026 for details.
Download the full final newsletter [here]