IBM targets quantum advantage by 2026

“Our mission at IBM is very simple: build useful quantum computing, bring it to the world, and make the world 1% better,” Chow stated.

The road to quantum utility

Quantum mechanics, now over a century old, has long provided the framework for understanding the physical world, but Chow noted that classical computers have so far been unable to simulate quantum systems at scale.

“It is the operating system of the universe,” he said, “but in many ways, the computers we’ve built – our phones, laptops, traditional servers – haven’t been able to run that maths at scale. That’s what quantum computing is starting to unlock.”

This potential spans applications in pharmaceuticals, materials science, financial modelling, and food security – all billion-dollar industries. Boston Consulting Group has projected a $500 billion market opportunity for quantum technologies at maturity, a figure Chow used to emphasise the scale of the opportunity.

Hardware progress: scaling with quality

IBM has consistently focused on scaling quantum hardware through both qubit count and quality, measured by error rates and circuit depth. The company has taken a manufacturing-led approach, leveraging its background in semiconductor fabrication.

“We’re not approaching this as a research project,” Chow said. “We’re doing cycles of learning, year over year, going from node to node, as we did in semiconductor development.”

In 2022, IBM surpassed the 100-qubit mark with its Eagle processor – a symbolic milestone because 100 qubits cannot be simulated using even the most powerful classical supercomputers due to exponential memory requirements. This was followed in 2023 by the demonstration of quantum utility using a 127-qubit processor.

Quantum utility, according to Chow, is defined by a quantum computer’s ability to run accurate computations beyond the reach of brute-force classical simulations. IBM validated this by solving a particular spin model with no known exact classical solution. While several approximate classical solvers attempted to replicate the result, their answers diverged significantly.

Software and speed: the Qiskit edge

Alongside hardware development, IBM has also prioritised software tools to enable performance and openness. Its open-source software development kit, Qiskit, remains the most widely used quantum SDK, with a 74% adoption rate according to the 2024 Unitary Foundation survey.

Chow highlighted the importance of performance benchmarking: “In the early days, Qiskit was designed to work well at the five-qubit level. Today, it is tuned for 100-plus qubit systems. You don’t want your software to be the bottleneck.”

With Qiskit 2.0, IBM claims major improvements in circuit construction efficiency and runtime speed, verified through an open-source suite of benchmarks. Chow stressed that this software layer is critical for unlocking hardware value, especially as users scale up to real workloads involving many iterations of quantum and classical computations.

A growing ecosystem

Since putting its first quantum systems on the Cloud, IBM has deployed 80 quantum computers globally and engaged over 600,000 users. Through its learning platform, the company has reached 10 million learners, and its IBM Quantum Network now includes over 275 organisations – spanning industry, national labs, universities, and startups.

Chow framed this as an accelerating cycle: “Our cycles of learning are starting to accelerate. We’re seeing real momentum now.”

Looking ahead, IBM plans to launch a new processor later in 2025 – IBM Quantum Condor – featuring under 220 qubits in a square lattice layout, designed to reduce circuit depth and improve error rates. It builds on the Heron class of tunable-coupler processors, which Chow said delivered a 5× improvement in error rates over previous designs.

This trajectory underpins IBM’s confidence in achieving quantum advantage by 2026 – defined as the point where quantum computing outperforms classical systems in efficiency, accuracy, and cost-effectiveness.

“Quantum-centric supercomputing will be the architecture that gets us there,” said Chow. “It’s a hybrid model, using the best of classical and quantum.”

Already, IBM has been collaborating with partners to simulate real molecular systems that lie beyond the reach of classical methods.

Beyond quantum advantage, IBM is targeting fault-tolerant quantum computing – scalable systems that can reliably perform long-duration computations. Chow hinted at forthcoming announcements related to IBM’s approach using quantum low-density parity-check (LDPC) codes, and a roadmap towards building those future systems. 

He concluded with a call to action: “Get started with quantum, because it’s already here. You can access machines now that are exploring utility-scale problems.”