IBM Targets 2025 for Quantum Processor Built to Tackle Real Workloads
IBM Quantum Nighthawk aims to compress circuits and reach 5,000-gate performance, bringing quantum closer to industry-grade applications
IBM is preparing to launch a new quantum processor designed to solve real-world problems at scale.
Called IBM Quantum Nighthawk, the chip will feature 120 square-lattice qubits and support quantum circuits with up to 5,000 two-qubit gates, enabling faster, denser calculations than previous architectures. Nighthawk's release in late 2025 represents IBM’s next leap toward commercial quantum computing.
“We’re really excited to bring Nighthawk to our users,” said Jerry Chow, IBM fellow and director of quantum systems at IBM Quantum. “It continues with our tunable coupler architecture and uses a square lattice instead of the heavy-hex layout. That allows us to compress circuit depths and make the execution of more complex workloads viable.”
The processor is part of IBM’s broader roadmap to achieve quantum advantage, when a quantum system outperforms classical methods in efficiency, cost, or accuracy.
Nighthawk builds on the momentum of IBM’s previous-generation processors, like Eagle and Heron, while introducing architectural innovations to scale performance with practicality.
“It’s not just about adding more qubits,” Chow said. “It’s about enabling useful computation.”
Inside Heron and Eagle
Before Nighthawk, IBM crossed a major threshold with two processors that marked quantum computing’s movement from theory to utility: Eagle and Heron.
Unveiled in 2021, IBM Eagle was the first quantum processor to surpass the 100-qubit barrier, featuring 127 qubits.
“It was the point where quantum states became too complex to be simulated even by the most advanced classical supercomputers,” Chow explained.
In computational terms, that meant IBM had entered a regime where classical simulation was no longer a viable fallback, moving into what many researchers call “quantum supremacy” or “quantum utility” territory.
Eagle featured a heavy-hex lattice architecture and advanced 3D packaging techniques such as bump bonds and multi-level wiring. These innovations not only stabilized the fragile quantum states but also paved the way for scalable hardware manufacturing.
“Producing Eagle on our timeline was possible because of IBM’s legacy in semiconductor engineering,” Chow said. “That lets us bring key technologies into quantum fabrication.”
But Eagle was only the beginning. In 2023, IBM launched Heron, a 156-qubit processor with a tunable coupler design that significantly improved performance metrics.
“Heron is the most performant quantum processor in the world today,” Chow said. “It enabled us to run 5,000 two-qubit gate circuits with dramatically lower error rates and reduced crosstalk.”
This leap in quality translated into measurable outcomes. Heron allowed IBM to demonstrate quantum utility, which is defined as the ability to solve problems that are out of reach for classical brute-force computation.
In one benchmark, IBM ran a simulation of an Ising spin model, a type of physics problem with no known exact classical solution.
“We published the result in Nature, and within weeks, multiple classical solvers attempted to replicate it,” Chow said. “But they all differed from each other in their results by about 20%. At that point, the quantum result becomes the arbiter.”
Importantly, Heron’s performance wasn’t confined to research labs. Five Heron processors have already been deployed via IBM’s cloud-accessible platform, with a sixth coming online soon. These are available for businesses, researchers, and developers to test directly.
“They're not just claims,” Chow emphasized. “They’re benchmarkable, usable, and accessible today.”
Alongside Heron came IBM’s latest release of Qiskit 2.0, the most widely used quantum software development kit. Optimized for 100+ qubit systems, Qiskit 2.0 reduces gate counts and accelerates quantum circuit compilation, avoiding performance bottlenecks.
“You definitely don’t want your software to be the limiting factor,” Chow said.
Heron’s architectural refinements, especially its coherence times in the 300-microsecond range and its best-case two-qubit gate error rates near 5e-4, establish a high-performance baseline on which Nighthawk will build. Unlike Eagle’s heavy-hex layout, Nighthawk’s square-lattice architecture is designed for even greater circuit compression, offering more efficient execution of complex algorithms.
Building at Engineering Scale
IBM is not just chasing quantum milestones—it is manufacturing them. Chow made clear that IBM’s approach is distinct from others in the field.
“We're the only team really approaching this not as a research project, but as an engineering challenge,” he said. That mindset is reflected in the company's track record.
IBM has deployed 80 quantum computers, trained 10 million learners, and built an active user base of over 600,000 developers.
At the heart of IBM’s hardware success is its legacy in semiconductor fabrication.
“We’re bringing technologies to quantum that are well established in classical electronics,” Chow explained. “Things like bump bonds, multi-level wiring, and through-silicon vias let us build reliably at scale.”
Manufacturing reliability and performance advances in coherence and gate fidelity help IBM offer devices capable of meaningful computations today.
“We’ve shown year-over-year improvements, not only in increasing qubit counts, but in making those qubits more stable and usable,” said Chow.
He noted that IBM’s superconducting qubit architecture offers dramatic speed advantages—up to hundreds of thousands of times faster than atomic or ion-based systems. That performance translates directly into cost efficiency for users running workloads on the IBM Quantum Platform.
From Advantage to Fault Tolerance
Speaking in London at The Economist's Commercializing Quantum Global 2025 conference on May 13, Chow said that quantum advantage—where quantum systems outperform classical ones in practical scenarios—is within reach.
“We believe we’ll achieve quantum advantage in 2026,” he said. The roadmap to that goal is based on “quantum-centric supercomputing,” a hybrid architecture that combines quantum and classical hardware to solve high-value problems.
IBM is already running simulations of molecular structures in collaboration with Japan’s RIKEN research institute, which challenge the limits of classical methods.
“We worked on modeling iron-sulfur clusters,” said Chow. “First on a solvable system, then on one that’s beyond exact classical simulation. The hybrid quantum-HPC method is now our reference point for comparison.”
Beyond 2026, IBM is set on the next major leap: fault-tolerant quantum computing. These systems, protected by advanced error-correcting codes, would allow sustained, reliable computation even in noisy environments.
“We want to get toward fault tolerance,” he said. “We’ll use low-overhead, scalable architectures based on our Dead City parity check codes.”
Fault-tolerant quantum systems could eventually tackle problems that remain intractable for even the most powerful classical supercomputers—optimizing global supply chains, designing new pharmaceuticals, or modeling complex climate systems in unprecedented detail.
IBM’s upcoming architecture announcement is expected to shed more light on how its future systems will scale to thousands—or even millions—of reliable qubits. Chow hinted that this next phase will combine hardware innovation with error-correcting software and classical orchestration, aiming to reach fully fault-tolerant systems by the end of the decade.
“Access to utility-scale quantum processors is already possible today,” he said, “but we’re now engineering for long-term resilience. The systems we’re building next won’t just be powerful—they’ll be dependable.”
As IBM prepares to launch Nighthawk and crosses the threshold toward quantum advantage, the focus is no longer just on performance metrics. It’s about designing a computing infrastructure that can stand the test of time, and finally put quantum systems to work across industries that have waited decades for their potential to become reality.