IBM, together with the University of Chicago, Algorithmiq, and Qedma, has completed a demonstration of 70 logical qubits on the Quantum Heron R3 superconducting system. The experiment executed 2,415 logical two-qubit gates and 468 logical T gates, finishing the task in roughly 15 minutes—beyond what top classical simulations can achieve. IBM's official announcement came on July 29, with the results reported in early August by outlets including Zhide Tech and NetEase, alongside arXiv preprints.
Key Points
- The core innovation is the space-time code, not the qubit count. Error detection is embedded directly inside the circuit, so a computation claimed to exceed classical limits carries built-in, statistically checkable confidence for the first time.
- It answers quantum computing's most awkward question: "How do you prove you actually beat classical?" Past quantum advantage experiments often devolved into disputes over correctness and possible classical shortcuts, since outsiders could not verify results. Space-time codes turn error detection into part of the circuit itself, dissolving the deadlock of "unprovable quantum advantage."
- This is a qualitative shift from "running it successfully" to "running it in a way that can be trusted."
- 468 logical T gates matter: logical T gates are the hard currency of universal fault-tolerant quantum computing. Running them stably at scale means a truly programmable, fault-tolerant machine is closer.
- The 15-minute runtime versus classical infeasibility suggests the window for "practical advantage" is narrowing. Combined with IBM's Starling roadmap (2029), the 70-logical-qubit result functions as a "credible milestone" for calibrating industry expectations of when quantum computing will actually be usable.
One-Line Takeaway
Quantum computing is no longer just about who has more qubits, but about whose "victory" can be proven—space-time codes turn a 70-logical-qubit advantage from a claim into a verifiable conclusion, a critical leap in engineering trustworthiness.