Beyond 120 Qubits: The Powerful IBM, AMD, and Oak Ridge Partnership Reshaping Quantum Computing

9 hours ago

11 min read

Quantum computing is approaching a critical stage in its development. Researchers have demonstrated increasingly sophisticated quantum operations, protected quantum information against noise, and combined quantum processors with classical supercomputers to investigate scientific problems. Yet turning these achievements into reliable, large-scale machines remains one of the most difficult challenges in modern computing.

A new collaboration between IBM, AMD, and the U.S. Department of Energy-funded Quantum Science Center at Oak Ridge National Laboratory (ORNL) is designed to address that challenge. Announced on November 5, 2026, the initiative introduces a co-design platform that brings quantum processors and high-performance classical computing together to accelerate research into quantum error correction.

The initial deployment connects IBM quantum computers, including its 120-qubit Quantum Nighthawk processor, with Oak Ridge's Umbra high-performance computing system. Umbra is powered by AMD Instinct MI355X GPUs and fifth-generation AMD EPYC CPUs, providing the classical processing resources required to investigate error correction, decoding algorithms, and software orchestration.

The collaboration aligns with the Department of Energy's Quantum Genesis Mission, which has established an ambitious objective: achieving 100 logical qubits capable of executing 100,000 hard operations by late 2028.

The significance extends beyond the development of another quantum processor. It reflects a growing understanding that practical quantum computing will depend on the coordinated operation of quantum hardware, classical supercomputers, advanced algorithms, and intelligent software systems.

Why Quantum Error Correction Is the Biggest Challenge in Quantum Computing

Quantum computers process information using quantum bits, or qubits. Unlike conventional bits, which represent either zero or one, qubits can exist in quantum superpositions and become correlated through entanglement.

These properties provide the foundation for quantum algorithms that may offer advantages in areas such as molecular simulation, materials science, optimization, and cryptography.

However, quantum states are extremely sensitive to environmental disturbances. Interactions with surrounding systems can introduce errors through processes such as decoherence, imperfect gate operations, control inaccuracies, and measurement noise.

A classical computer can generally copy and inspect information without fundamentally changing its representation. Quantum information cannot be protected through the same straightforward approach because unknown quantum states cannot simply be copied.

Quantum error correction addresses this problem by distributing logical information across multiple physical qubits. Carefully designed measurements reveal information about errors without directly measuring and destroying the encoded logical state.

These measurements generate information known as error syndromes. A decoder processes that information to estimate which errors occurred and determine the appropriate correction.

The distinction between physical and logical qubits is fundamental:

Physical qubits: The actual quantum hardware elements used to store and manipulate quantum information.
Logical qubits: Protected units of quantum information encoded across multiple physical qubits.
Error syndromes: Measurement outcomes that provide information about errors affecting the encoded quantum state.
Decoders: Classical algorithms that interpret syndrome data and identify likely error patterns.
Fault-tolerant operations: Computations designed to remain reliable even when individual hardware components occasionally make errors.

The objective is not to eliminate every physical error. Instead, quantum error correction aims to suppress the probability of logical failure sufficiently that increasingly complex computations can be performed reliably.

This creates a substantial computational challenge. As quantum systems grow, the classical processing required to interpret measurements and coordinate corrections can become increasingly demanding.

IBM, AMD, and Oak Ridge Build a Quantum-Centric Supercomputing Platform

The new platform brings together three distinct areas of technological expertise.

IBM contributes quantum processors, quantum control capabilities, and software interfaces that allow researchers to investigate error correction at a detailed hardware level.

Oak Ridge National Laboratory provides its scientific computing environment and expertise through the Quantum Science Center, which is funded by the U.S. Department of Energy.

AMD supplies the classical computing foundation through its Instinct MI355X GPUs and fifth-generation EPYC CPUs, integrated into Oak Ridge's Umbra system.

The collaboration is built around co-design, an approach in which hardware, software, algorithms, and system architecture are developed and evaluated together rather than independently.

This matters because an error correction algorithm cannot be evaluated solely by its mathematical properties. Researchers must also determine whether it can process real measurements quickly enough, operate within available computing resources, and remain effective under the limitations of actual quantum hardware.

The initial platform creates an environment in which these questions can be investigated using connected quantum and classical systems.

Technology component	Primary role
IBM Quantum Nighthawk	Executes quantum circuits and supports error correction experiments
IBM Quantum System API	Provides detailed access to quantum control and measurement capabilities
AMD Instinct MI355X GPUs	Accelerates computationally intensive classical workloads
Fifth-generation AMD EPYC CPUs	Supports general-purpose processing and system coordination
ORNL Umbra HPC system	Provides the classical high-performance computing environment
Quantum Science Center	Coordinates research into scalable quantum computing methods

The platform will also connect with the Oak Ridge Leadership Computing Facility in Tennessee, creating opportunities to investigate quantum computing as part of a wider scientific and AI computing ecosystem.

The 2028 Quantum Genesis Mission: Closing the Gap to 100,000 Hard Operations

One of the most important aspects of the partnership is its connection to the Department of Energy's Quantum Genesis Mission.

Through its Q Competition, the mission targets a quantum computer capable of operating with 100 logical qubits and executing 100,000 hard operations by late 2028.

The reported IBM research baseline illustrates the scale of the remaining challenge.

IBM has demonstrated experiments involving up to 70 logical qubits and 438 hard operations. Compared with the 100,000-operation target, the hard-operation count must increase by approximately 230 times.

Measurement	Current reported IBM demonstration	DOE target
Logical qubits	Up to 70	100
Hard operations	438	100,000
Required increase in hard operations	Baseline	Approximately 230-fold

These figures highlight why quantum computing progress cannot be measured simply by counting physical qubits.

A processor may contain more physical qubits but still struggle to execute long computations reliably. The ability to maintain logical information and perform repeated operations without accumulating unacceptable errors is a more meaningful measure of practical capability.

The 2028 target therefore represents a combined challenge involving quantum hardware quality, error correction efficiency, decoding speed, control systems, and software integration.

It is also important to distinguish a scientific target from a guaranteed delivery date. Achieving the mission's objective will require further experimental progress and successful integration across multiple technical layers.

Why Classical GPUs Are Essential to Fault-Tolerant Quantum Computing

Quantum computing is frequently presented as a potential successor to classical computing. In practice, the two technologies are more likely to operate together.

Quantum processors are specialized systems that may offer advantages for particular computational problems. Classical CPUs and GPUs remain essential for managing workloads, processing data, coordinating operations, and executing algorithms that do not require quantum hardware.

Error correction is an especially important example of this relationship.

Quantum processors generate measurement outcomes that must be interpreted by classical algorithms. As the number of physical qubits and the complexity of error correction codes increase, the volume and frequency of classical processing can also grow.

GPUs are particularly useful for workloads involving large numbers of parallel mathematical operations. Depending on the decoding algorithm, they can help process syndrome information, evaluate possible error patterns, and accelerate computationally intensive optimization tasks.

However, raw processing power is only part of the solution. Error correction systems must also satisfy strict timing requirements.

If the classical system takes too long to process measurements, it may become difficult to coordinate correction operations efficiently with the quantum processor.

This makes latency, communication architecture, memory movement, and software scheduling important considerations alongside GPU performance.

The IBM-AMD-ORNL collaboration provides an environment for testing how these components should be organized to support increasingly complex quantum systems.

Quantum Nighthawk and the Importance of Real-Time Quantum Control

IBM's 120-qubit Quantum Nighthawk processor is a central component of the initial deployment.

Its square-lattice connectivity, high circuit execution speed, and real-time pulse control provide researchers with capabilities for investigating different quantum error correction strategies.

Connectivity matters because quantum error correction codes depend on interactions between qubits. The physical arrangement of qubits and the available connections influence how efficiently operations can be executed.

A suitable architecture can reduce unnecessary operations and help researchers manage the complexity of error correction circuits.

Real-time pulse control provides another important advantage. Quantum processors are manipulated through precisely timed control signals. Researchers need detailed control over these signals to test how different procedures affect quantum states and error rates.

IBM's Quantum System API provides access to capabilities that include:

Measuring qubits during an ongoing calculation.
Responding to measurement results in real time.
Controlling the timing of signals used to protect quantum information.
Examining the operations executed by quantum hardware.
Accessing raw measurement information for detailed research.

These capabilities allow researchers to study error correction under conditions that more closely reflect the behavior of actual quantum systems.

The platform can also help compare different approaches according to their requirements for physical qubits, execution time, measurement samples, and classical processing resources.

This is essential for determining where accelerated computing can deliver the greatest benefit.

The Quantum Error Correction Methods Researchers Will Investigate

The collaboration is expected to explore several approaches that could contribute to more scalable quantum error correction.

Concatenated quantum error correction codes

Concatenated codes use multiple layers of encoding to protect logical information. An initial encoding can be combined with additional protection levels to reduce the probability of logical errors.

The challenge is that additional layers can increase the number of physical qubits and operations required. Researchers must therefore identify configurations that provide sufficient protection without making the overall system impractical.

Space-time codes

Space-time approaches use information collected across both the physical arrangement of qubits and successive measurement cycles.

By examining error patterns over time, these techniques can help identify how errors develop and propagate through a quantum computation.

Their effectiveness depends on the quality of measurement data and the ability of classical systems to process it efficiently.

Algorithm-aware decoders

Not every quantum computation has the same requirements. Algorithm-aware decoding takes the structure of the intended computation into account when interpreting error information.

This creates opportunities to optimize error correction for particular scientific workloads rather than relying exclusively on general-purpose decoding strategies.

Hardware-aware circuit optimization

Quantum processors have physical limitations involving connectivity, gate performance, control accuracy, and execution time.

Hardware-aware optimization adapts quantum circuits to the characteristics of the available processor. This can reduce unnecessary operations and improve the practical execution of error correction procedures.

Software orchestration

A fault-tolerant quantum system requires coordination between quantum processors, classical accelerators, control software, decoders, and workload management systems.

Software orchestration ensures that these components work together as an integrated computing environment.

The long-term objective is to bring these methods into a unified architecture capable of supporting increasingly complex quantum applications.

From Scientific Research to Materials Discovery and Fusion Energy

Quantum computing's long-term value may be especially significant in scientific research, where accurately modeling complex quantum systems is computationally difficult for conventional computers.

Molecules, magnetic materials, and advanced chemical structures are governed by quantum mechanical interactions. Classical simulation becomes increasingly demanding as the complexity of these systems grows.

A reliable quantum computer could eventually help researchers investigate molecular properties, chemical reactions, material behavior, and other problems that are difficult to model efficiently using conventional methods.

The IBM and ORNL collaboration already provides a foundation for this research direction.

Earlier work involving IBM quantum computers and Oak Ridge's Frontier supercomputer demonstrated how combined quantum and classical workloads could simulate real magnetic materials and calculate molecular configurations of a promising material associated with fusion fuel production.

These examples illustrate the potential of quantum-centric supercomputing. Rather than replacing classical supercomputers, quantum processors can be integrated into scientific workflows where their capabilities may complement existing computational methods.

Potential applications include:

Materials science: Investigating the electronic properties of complex materials.
Chemistry: Exploring molecular structures and reaction mechanisms.
Fusion research: Supporting the study of materials relevant to fusion energy systems.
Drug discovery: Improving the understanding of molecular interactions and chemical behavior.
Energy technology: Investigating materials for batteries, catalysts, and advanced energy systems.

These applications remain dependent on achieving reliable quantum hardware and demonstrating useful computational advantages for specific problems.

IBM and AMD's Existing Collaboration Creates a Foundation for Integration

The new initiative builds on previous work between IBM, Oak Ridge, and AMD.

IBM and ORNL collaborate through the Quantum Science Center, an ORNL-led research center established under the U.S. National Quantum Initiative.

IBM and AMD have also worked together to accelerate classical components of quantum-centric computing.

One example involves sample-based quantum diagonalization, or SQD, a workflow that combines quantum sampling with classical computational methods. AMD GPUs are being used to accelerate the classical diagonalization stage.

The companies also previously demonstrated an algorithm for decoding quantum errors on AMD hardware.

These efforts are important because the development of fault-tolerant quantum computing requires a broader ecosystem of compatible hardware and software.

A quantum processor alone cannot provide a complete solution. The surrounding classical infrastructure must be capable of handling the computational demands created by increasingly sophisticated quantum workloads.

The new platform provides an opportunity to explore how future AMD processors, networking technologies, and open software can evolve alongside quantum computing systems.

The Road Toward IBM Quantum Starling and Future Supercomputers

IBM's longer-term roadmap includes Quantum Starling, which the company describes as its first fault-tolerant quantum computer.

The new co-design platform may help inform the hardware architectures, error correction techniques, and software systems needed for that objective.

The collaboration also points toward a future in which quantum processors, AI accelerators, and high-performance computing systems operate as components of larger scientific computing platforms.

This could eventually lead to systems that combine:

Quantum processors for specialized quantum workloads.
GPUs for accelerated classical computation and AI processing.
CPUs for general-purpose execution and coordination.
Advanced networking for communication between computing resources.
AI-based software tools for optimization, monitoring, and workload management.
Fault-tolerant control systems for reliable quantum operations.

The reference to future Super Intelligence and HPC systems reflects an emerging vision of highly integrated computing environments. However, the practical development of these systems will depend on demonstrated technical performance, reliability, and commercially or scientifically valuable applications.

Key Challenges on the Path to Practical Quantum Computing

Despite substantial progress, several challenges remain before fault-tolerant quantum computers can become practical scientific tools.

Challenge	Why it matters
Physical error rates	High error rates increase the resources required for logical protection
Logical qubit scaling	More reliable logical qubits are needed for complex computations
Decoder performance	Classical processing must keep pace with quantum measurements
Hardware connectivity	Physical architecture influences circuit complexity and execution efficiency
Control precision	Accurate pulse timing and measurement are necessary for reliable operations
System integration	Quantum, classical, networking, and software components must work together
Resource overhead	Error correction can require substantially more physical qubits than logical qubits

There is also a fundamental economic question. Fault-tolerant quantum systems may require extensive infrastructure, specialized engineering, and significant operational resources.

Their practical value will ultimately depend on whether they can solve meaningful problems more effectively than available classical alternatives.

For this reason, collaborations between national laboratories, semiconductor manufacturers, and quantum computing companies are strategically important. They combine research expertise with the hardware and infrastructure needed to evaluate complete computing systems.

Quantum Computing's Future Depends on the Integration of Quantum and Classical Power

The collaboration between IBM, AMD, and Oak Ridge National Laboratory represents an important development in the effort to make quantum computing reliable and scientifically useful.

By combining IBM's 120-qubit Nighthawk processor with AMD-powered high-performance computing and Oak Ridge's research infrastructure, the initiative addresses one of the central requirements of fault-tolerant quantum computing: scalable and efficient error correction.

The Department of Energy's objective of 100 logical qubits performing 100,000 hard operations by late 2028 provides a measurable milestone for the field. Reaching it will require substantial progress in decoding, hardware control, quantum architecture, and classical computing integration.

For Dr. Shahid Masood and the expert team at 1950.ai, this development illustrates a wider transformation in computing. The future may not be defined by a competition between quantum computers, classical supercomputers, and AI systems, but by the ability to integrate them into a unified technological ecosystem.

Key Takeaways

IBM, AMD, and Oak Ridge National Laboratory have introduced a co-design platform for quantum error correction.
The initial deployment combines IBM's 120-qubit Quantum Nighthawk with ORNL's Umbra HPC system.
Umbra uses AMD Instinct MI355X GPUs and fifth-generation AMD EPYC CPUs.
The initiative supports the DOE Quantum Genesis Mission's target of 100 logical qubits performing 100,000 hard operations by late 2028.
IBM's reported baseline includes up to 70 logical qubits and 438 hard operations, leaving an approximately 230-fold increase in hard-operation count to reach the target.
Classical GPUs can help accelerate decoding and other computationally demanding parts of quantum error correction.
Research will examine concatenated codes, space-time codes, algorithm-aware decoders, hardware-aware optimization, and software orchestration.
The collaboration could inform IBM's longer-term development of Quantum Starling and future quantum-centric supercomputing systems.

Future Outlook

The next phase of quantum computing will be determined by the ability to move beyond isolated demonstrations toward integrated systems that can maintain logical information and execute increasingly complex computations reliably.

The partnership between IBM, AMD, and Oak Ridge provides a framework for testing that integration. Its results could influence how future quantum processors are designed, how classical computing resources are deployed, and how scientific institutions approach computationally demanding research.

The ultimate milestone will not simply be a larger quantum processor. It will be a reliable computing system capable of delivering meaningful scientific results through the coordinated power of quantum hardware, classical supercomputing, and advanced software.