Oliver Dial of IBM: Quantum Advantage Is Happening This Year

What It Covers

IBM Quantum VP Oliver Dial explains where quantum computing stands in 2026, covering the distinction between quantum utility and quantum advantage, how 156-qubit superconducting processors work, why the new Gross error-correcting code reduces qubit overhead by 10x, and why fault-tolerant systems are now projected for 2029.

Key Questions Answered

• •Quantum Advantage Threshold: IBM's Quantum Advantage Tracker, a public GitHub-based leaderboard modeled on Hugging Face, allows researchers to post verified head-to-head comparisons of quantum versus classical performance on specific problems. Enterprises should monitor this tracker now, as several events have already been logged and verifiable advantage on real problems may already be within reach in 2026.
• •Qubit Count and Simulation Barrier: Classical computers cannot efficiently simulate quantum processors once qubit counts exceed roughly 50–100. IBM's current flagship Heron processor runs 156 physical qubits, surpassing that threshold. The 2023 Condor device reached 1,000 qubits but was decommissioned immediately due to error rates too high to be computationally useful, illustrating that raw qubit count alone is insufficient.
• •Gross Code Efficiency Breakthrough: Previous error-correcting codes required approximately 100 physical qubits per one logical qubit, making fault-tolerant systems impractical. IBM's newly developed Gross code, co-designed around hardware constraints of six connections per qubit and wire reach of roughly 10 qubits, achieves an order-of-magnitude improvement in efficiency, fundamentally reshaping the 2029 fault-tolerant timeline.
• •Nighthawk Processor Architecture: IBM's 2026 Nighthawk processor uses 120 qubits but adds a fourth coupler connection per qubit, up from three on previous devices. This connectivity increase allows more efficient gate operations without routing swap operations across the chip, improving effective computational depth even at a slightly reduced physical qubit count compared to the 156-qubit Heron.
• •Enterprise Workforce Preparation: Organizations operating on four-to-five-year investment horizons should begin training staff in quantum algorithm mapping now. Finding people who understand both domain-specific problems and quantum computing well enough to map one onto the other takes years. Near-term heuristic algorithms in optimization and chemistry are approaching practical thresholds, making preparation relevant before fault-tolerant systems arrive in 2029.