Industry News

IBM Unveils Next-Gen 133-Qubit ‘Heron’ Quantum Processor

Yorktown Heights, N.Y., USA (Nov 2024) – IBM has announced a new superconducting quantum processor, code-named “Heron,” featuring 133 qubits and a host of architectural advances. The IBM Quantum Heron chip was unveiled at the IBM Quantum Summit 2023 as the company’s latest milestone in its quantum computing roadmap. IBM touts Heron as a next-generation processor that delivers significantly improved performance and reliability compared to its predecessors. This 133-qubit device introduces new technologies aimed at boosting quantum computation capability while laying the groundwork for IBM’s future quantum systems.

Key Features of the Heron Chip

133 Superconducting Qubits: Heron contains 133 qubits, slightly more than IBM’s previous 127-qubit flagship (Eagle) and built using fixed-frequency transmon qubits. It surpasses the 100-qubit scale that IBM first achieved in 2021 with Eagle, marking another step in qubit count while maintaining stability.

Tunable Coupler Architecture: A standout innovation in Heron is the use of tunable couplers between qubits. Unlike earlier processors with static coupling, Heron’s couplers can be adjusted to control interactions, which virtually eliminates cross-talk (undesired interference between neighboring qubits). This means qubits can be better isolated when not actively interacting, leading to cleaner operations.

Improved Fidelity and Gate Performance: Thanks to its new architecture, Heron achieves a 3–5× improvement in overall device performance relative to IBM’s 127-qubit Eagle processor. In practical terms, IBM reports Heron can execute roughly 1,800 quantum gates within a single coherence cycle, about four times the number of gate operations Eagle could run in the same period. This makes Heron IBM’s lowest-error, highest-performing processor to date, significantly reducing error rates and increasing the complexity of circuits that can be run reliably.

Foundation for Future Systems: IBM engineers designed Heron with a vision toward scalable quantum-centric supercomputing. It is built to integrate into IBM’s new Quantum System Two infrastructure as a module that can work in tandem with other chips. IBM’s roadmap positions Heron as a foundational building block for larger systems – in fact, the first IBM Quantum System Two is configured to house three Heron processors working together in one cryogenic setup. This modular approach will allow IBM to link multiple Heron chips for increased qubit counts and parallel quantum operations, without sacrificing fidelity.

Why Heron Is a Significant Milestone

IBM’s unveiling of the Heron chip represents an important strategic step in the company’s quantum timeline. Rather than only increasing qubit numbers, Heron emphasizes quantum computing “quality” and utility. By drastically lowering error rates and enabling deeper circuits, the Heron processor pushes IBM closer to the era of practical quantum computing. IBM has dubbed this the era of “quantum utility,” where quantum computers begin to solve meaningful problems that are intractable for classical machines. The Heron’s improved performance and stability are key enablers for that goal, allowing researchers and developers to run more complex algorithms with confidence in the results.

IBM envisions Heron as the backbone of its near-term quantum offerings. “With Heron, we have developed a qubit and gate technology that we’re confident will form the foundation of our hardware roadmap going forward,” said Jay Gambetta, IBM Fellow and VP of Quantum, in an IBM blog post. The chip’s introduction goes hand-in-hand with IBM’s broader strategy to scale up quantum systems in a modular way. By installing multiple Heron processors in the IBM Quantum System Two and connecting them, IBM aims to scale to larger effective qubit counts while maintaining low error rates – a different approach from simply building one enormous chip. This modular, quantum-centric supercomputing vision is seen as crucial for tackling practical applications and eventually implementing error-corrected quantum computing in the coming years.

The timing of Heron’s debut also aligns with IBM’s extended quantum roadmap shared at the 2023 Quantum Summit. Alongside Heron, IBM revealed progress toward ultra-large processors (for example, a 1,121-qubit chip named “Condor” was showcased as a leap in scale). Heron complements these developments by addressing the challenge of quantum error rates: while Condor and similar chips push the frontier of qubit count, Heron’s design improves the coherence and fidelity needed to make those qubits useful. This dual focus on scale and quality marks a pivotal shift in IBM’s quest for quantum advantage, bringing the industry closer to quantum systems that can handle real-world computational tasks.

How Heron Differs from IBM’s Earlier Chips (Eagle and Falcon)

Heron vs. Eagle (127 Qubits, 2021): The previous IBM flagship, Eagle, was introduced in 2021 with 127 qubits as IBM’s first processor to cross the 100-qubit threshold. Eagle employed IBM’s “heavy-hex” qubit layout – a hexagonal qubit arrangement that connects each qubit to only 2 or 3 neighbors – which was designed to reduce errors from unwanted interactions. Eagle also pioneered new 3D packaging techniques (such as multi-level wiring and readout multiplexing) to manage control wiring for so many qubits. This was a major leap in scale: Eagle made quantum circuits complex enough that they could no longer be reliably simulated on classical supercomputers, heralding a milestone in quantum capability.

Where Heron builds on Eagle is in refining the processor’s coherence and gate fidelity rather than just increasing qubit count. Heron’s qubit count (133) is only slightly above Eagle’s, but thanks to the tunable couplers and improved design, Heron can run far more gates within the qubits’ coherence time (four times Eagle’s gate depth). Essentially, Heron can execute longer quantum computations before errors accumulate, whereas Eagle was more limited by cross-talk and noise. IBM reports a 3–5× performance boost with Heron over Eagle, reflecting improvements in error rates and consistent operation. In practice, this means Heron is much less error-prone than Eagle and can handle more complex algorithms or error mitigation techniques without decohering. The heavy-hex connectivity is preserved in Heron, but the addition of tunable couplers allows engineers to turn off interactions when qubits aren’t actively engaged, a capability Eagle lacked. As a result, Heron nearly eliminates the residual coupling errors (“spectator” errors) that even Eagle’s heavy-hex layout couldn’t fully avoid. In short, Eagle was about achieving scale, whereas Heron is about achieving stable, high-fidelity computation at scale.

Heron vs. Falcon (27 Qubits, 2019): Falcon was IBM’s earlier-generation processor (deployed in 2019) with 27 qubits, which introduced the heavy-hexagonal lattice architecture as a means to improve reliability. In the Falcon design, limiting each qubit’s neighbors helped decrease error rates caused by frequency collisions and cross-talk, yielding more stable multi-qubit operations for its size. Falcon and its successor, the 65-qubit Hummingbird (2020), demonstrated IBM’s ability to scale up qubits while keeping errors in check using clever layout and fabrication techniques. However, those earlier chips still employed fixed coupling between qubits and faced challenges as qubit counts grew.

Comparatively, Heron’s 133 qubits represent a five-fold increase in qubit count over Falcon, but more importantly, Heron embodies the evolution of IBM’s technology to maintain qubit quality at a higher scale. The tunable coupler approach in Heron is a departure from Falcon’s fixed couplers, directly tackling the cross-talk and connectivity trade-offs. IBM’s internal metrics illustrate this progress: for example, Falcon’s heavy-hex layout provided a solid foundation with fewer neighbor-induced errors, but Heron’s design pushes error rates down even further to unprecedented lows. Additionally, Heron is built to operate as part of a larger multi-chip system (something not possible with Falcon’s era of single-chip systems), reflecting how IBM’s strategy has shifted toward modular expansion. In essence, Falcon set the stage with a novel qubit layout for error reduction, Eagle scaled that layout to triple-digit qubit counts, and now Heron refines the architecture with tunable inter-qubit coupling and modular integration. This progression underscores IBM’s iterative approach: each generation addresses the limitations of the previous one – from Falcon’s connectivity trade-offs to Eagle’s wiring and layout challenges, and now to Heron’s focus on error suppression and scalability.

Context in IBM’s Quantum Roadmap

Heron’s debut fits into IBM’s broader quantum roadmap as of late 2023, which balances increasing qubit counts with improving quantum computing “quality” (coherence, fidelity, and useful circuit depth). Over the past few years, IBM has steadily advanced from the 20–50 qubit scale to over 100 qubits: the 27-qubit Falcon in 2019 was followed by the 65-qubit Hummingbird in 2020 and the 127-qubit Eagle in 2021. Each of these milestones expanded the scale while introducing design innovations to manage error rates. In 2022, IBM unveiled the 433-qubit Osprey processor (pushing the scale further, though still based on the same fundamental architecture). By 2023, IBM’s roadmap reached the “Quantum Summit” moment – delivering on the promise of a 1,121-qubit chip (Condor) and simultaneously introducing Heron’s new architecture for reliability. This dual achievement – one chip emphasizing sheer scale, and another emphasizing stability and modularity – reflects IBM’s two-pronged strategy to eventually achieve quantum advantage (solving useful problems faster than classical computers).

Importantly, IBM’s Quantum System Two (revealed conceptually alongside Eagle and now realized with Heron processors) is central to the roadmap at this stage. System Two is a next-generation cryogenic system designed to link multiple quantum chips via high-density interconnects, effectively allowing quantum processors to grow beyond the size of a single wafer. In December 2023, IBM reported that Quantum System Two is being equipped with three Heron chips as a prototype of this modular quantum computing approach. This places Heron at the heart of IBM’s near-term strategy: rather than waiting for one monolithic thousand-qubit device to be error-free, IBM is leveraging Heron to build “utility-scale” quantum systems by networking manageable 133-qubit modules. As IBM’s roadmap indicated at the Summit, the goal over the next several years is to scale up both qubit count and qubit quality in tandem. Heron is a key enabler of that vision, bridging the gap between early quantum hardware and the long-term aim of fault-tolerant quantum computers.

In summary, the IBM Heron chip announcement marks a pivotal moment in 2023 for quantum tech watchers. It reinforces that progress in quantum computing is not just about making chips bigger, but making them better. With 133 qubits, tunable couplers, and significantly reduced errors, IBM’s Heron processor is poised to accelerate the journey toward practical quantum computing applications. It builds on the legacy of Falcon and Eagle while steering IBM’s hardware roadmap toward more scalable and usable quantum machines. Tech-savvy observers view Heron as a confirmation that IBM is entering the next phase of quantum development – one where quantum processors inch closer to real-world utility by combining scale with reliability, a crucial combination for the coming era of quantum computing.

Marin Ivezic

I am the Founder of Applied Quantum (AppliedQuantum.com), a research-driven professional services firm dedicated to helping organizations unlock the transformative power of quantum technologies. Alongside leading its specialized service, Secure Quantum (SecureQuantum.com)—focused on quantum resilience and post-quantum cryptography—I also invest in cutting-edge quantum ventures through Quantum.Partners. Currently, I’m completing a PhD in Quantum Computing and authoring an upcoming book “Practical Quantum Resistance” (QuantumResistance.com) while regularly sharing news and insights on quantum computing and quantum security at PostQuantum.com. I’m primarily a cybersecurity and tech risk expert with more than three decades of experience, particularly in critical infrastructure cyber protection. That focus drew me into quantum computing in the early 2000s, and I’ve been captivated by its opportunities and risks ever since. So my experience in quantum tech stretches back decades, having previously founded Boston Photonics and PQ Defense where I engaged in quantum-related R&D well before the field’s mainstream emergence. Today, with quantum computing finally on the horizon, I’ve returned to a 100% focus on quantum technology and its associated risks—drawing on my quantum and AI background, decades of cybersecurity expertise, and experience overseeing major technology transformations—all to help organizations and nations safeguard themselves against quantum threats and capitalize on quantum-driven opportunities.
Share via
Copy link
Powered by Social Snap