IBM and AMD recently announced a strategic collaboration to develop quantum-centric supercomputing architectures that combine quantum computers with high-performance computing infrastructure. The partnership is based on a memorandum of understanding between the companies, with no immediate financial exchange.

The partnership focuses on creating hybrid classical-quantum computing platforms that leverage IBM’s leadership in quantum computing and AMD’s high-performance computing capabilities.

The collaboration aims to address a fundamental requirement of fault-tolerant quantum computing: the need for substantial classical computing resources to maintain quantum state, perform error correction, and execute hybrid algorithms.

AMD will provide CPUs, GPUs, and FPGAs to integrate with IBM’s quantum systems, while both companies will work on developing new algorithms that require both quantum and classical processing paradigms.

Technical Details

A quantum-centric supercomputing architecture uses quantum computers as specialized processing units within a broader high-performance computing infrastructure.

IBM’s approach distributes computational workloads based on optimal processing decisions, with quantum computers handling specific calculations, such as atomic and molecular simulations. At the same time, classical supercomputers manage large-scale data analysis tasks.

IBM’s current quantum systems already incorporate classical computing infrastructure for basic operations. The IBM Quantum System Two, IBM’s first modular quantum computer, utilizes AMD’s Xilinx FPGAs equipped with RF modulators to manipulate qubit states through microwave pulse generation.

These FPGAs serve as critical interface components between classical control systems and quantum processing units.

Hardware Integration Points

AMD’s contribution will center on three primary hardware categories within IBM’s quantum infrastructure:

• The company’s EPYC CPUs will provide computational power for pre-processing and post-processing tasks within quantum workflows.
• AMD Instinct GPUs will handle data conditioning before quantum processing and AI-driven analysis of quantum results.
• The integration also incorporates AMD’s Xilinx FPGA portfolio, building on existing quantum control infrastructure already deployed in IBM systems.

Quantum Scale and Limitations

IBM’s current Quantum Heron systems operate with 156 qubits, exceeding the simulation capabilities of any publicly known classical supercomputer.

This level of capability marks a transition point at which quantum systems begin to demonstrate computational advantages over classical simulation approaches.

The partnership targets IBM’s goal of delivering fault-tolerant quantum computers by 2030. AMD’s real-time processing capabilities, through its FPGA and GPU portfolio, will provide the essential error correction infrastructure required for fault-tolerant quantum operations.

Analysis

The IBM-AMD quantum computing partnership takes a pragmatic approach to addressing the fundamental infrastructure requirements of fault-tolerant quantum computing, with both companies collaborating.

The collaboration creates a vertically integrated quantum-classical computing stack that differentiates both companies from competitors pursuing purely quantum or purely classical approaches.

IBM gains access to leading-edge classical computing hardware without developing internal capabilities, while AMD positions itself within the emerging quantum computing ecosystem without making direct investments in quantum hardware development.

AMD benefits from its association with advances in quantum computing while leveraging its existing strengths in high-performance computing. The company’s successful transformation from primarily a CPU supplier to a comprehensive data center computing provider gains additional validation through the development of quantum computing.

The partnership’s success depends on several critical factors, including the development of efficient quantum-classical algorithms, the achievement of fault-tolerant quantum operations, and the demonstration of clear computational advantages over purely classical approaches. While current performance metrics show promise in specific applications, broader market adoption requires continued technical advancement and cost reduction.

The collaboration between these two industry leaders provides a credible path toward practical quantum computing applications. This could be a significant step forward in the evolution from experimental quantum research to production-ready quantum-enhanced computing solutions.

It’s very early days for the quantum computing market, with significant gains not expected until close to the end of the decade. The timeline doesn’t detract from the engagement’s substantial value. After all, the early winners in the current AI boom are the result of decisions made a decade or more ago.

Transformational technologies, such as quantum computing, require a long gestation period. That makes collaborations like the one between IBM and AMD potentially significant, and one we’ll continue to watch.

Competitive Outlook & Advice to IT Buyers

IBM’s quantum leadership faces competitive pressure from Google, IonQ, Rigetti, Quantinuum, and other quantum computing developers…

These sections are only available to NAND Research clients and IT Advisory Members. Please reach out to [email protected] to learn more.