IBM's 70-Qubit Computer Solved What Supercomputers Can't

In a result that researchers have been working toward for decades, IBM and scientists at the University of Chicago have completed a quantum computation that leading classical supercomputers could not practically reproduce — and they did it in roughly 15 minutes.

The system used 70 error-corrected logical qubits, a landmark achievement in the field of quantum computing. The result, published in late August 2026, not only demonstrates what the field calls "quantum advantage" — the point at which quantum machines can do things classical computers cannot — but also provides statistical evidence that the output was correct, addressing one of the central challenges of verifying quantum results.

This is a milestone the field has been chasing for years. Quantum computers work by exploiting the principles of quantum mechanics: their fundamental units, called qubits, can exist in superpositions of 0 and 1 simultaneously, and groups of qubits can become "entangled" in ways that allow them to perform certain calculations exponentially faster than classical systems. But qubits are fragile. They are easily disrupted by heat, vibration, and electromagnetic noise — a phenomenon called decoherence — which introduces errors that cascade through a computation.

The solution is quantum error correction, a technique that uses multiple physical qubits to encode a single protected "logical" qubit. This overhead is enormous: current error correction schemes can require hundreds or thousands of physical qubits to reliably protect one logical qubit. The 70 logical qubits in this experiment represent a significant step toward the scale needed for practical, real-world applications.

What makes the result especially significant is the nature of the problem solved. The team chose a computational task specifically designed to be hard for classical computers at scale — a class of physics simulation problems that grows exponentially more complex as the system size increases. At 70 logical qubits, the problem crossed a threshold where even the best classical methods, running on the most powerful supercomputers available, could not compete in a practical timeframe.

"This is the kind of milestone we've been working toward since the early days of the field," said one researcher involved in the project. The 15-minute computation time is itself striking — comparable to a typical lunch break, for a problem that might take a classical supercomputer far longer to approximate.

The implications extend well beyond academic physics. Quantum computers at this scale — and eventually at larger scales — are expected to have transformative applications across a range of industries. Drug discovery stands out: simulating the quantum behavior of molecules is exactly the kind of problem classical computers struggle with, and where quantum machines could accelerate the identification of new medicines. Materials science, financial modeling, and optimization problems in logistics and supply chains are other areas where quantum advantage could eventually translate into real-world impact.

For IBM, the result caps years of steady progress along its published quantum roadmap. The company has been systematically scaling its quantum systems, improving error rates, and building the software infrastructure to make quantum computing accessible to researchers and businesses. The 70-qubit error-corrected milestone represents a qualitative jump — not just more qubits, but more reliable ones.

The University of Chicago collaboration reflects a broader trend of academic-industry partnerships driving quantum progress. University researchers bring theoretical depth and the freedom to explore novel approaches; companies like IBM provide engineering scale, hardware resources, and the infrastructure to run computations at speeds that matter.

The field still has a long road ahead. Scaling from 70 logical qubits to the thousands or millions needed for some of the most ambitious applications will require continued breakthroughs in error correction, qubit quality, and system integration. But the September 2026 result provides something the field has long needed: concrete, verifiable evidence that the quantum advantage is real and achievable today, not just promised for some future decade.

For anyone who has watched quantum computing develop from a purely theoretical curiosity into an operational technology, the 15 minutes it took IBM's machine to complete what supercomputers could not is a profound moment — the clearest proof yet that a new era of computation has arrived.