AMD’s expanded partnership with PsiQuantum reflects an important shift in the development of quantum computing. The challenge is no longer simply to build more capable quantum processors, but to create the classical computing, control, networking and error-correction infrastructure required to operate them at scale. The collaboration provides fresh evidence for the industrial transformation examined in The Quantum Space’s State of Quantum 2026 report, while leaving the question of sustainable commercial demand firmly open.
Quantum computing is increasingly becoming a systems-engineering challenge. After years in which attention focused on improving processors, increasing qubit counts and demonstrating new computing architectures, the industry is turning towards the infrastructure required to make those capabilities operate together.
That infrastructure extends well beyond the quantum processor itself. It includes conventional computing hardware, real-time control systems, error-correction technologies, networking, software and the industrial capabilities needed to manufacture and maintain complete machines. As quantum systems become more sophisticated, these supporting technologies are becoming central to their development.
On 12 November, AMD and PsiQuantum announced an expanded strategic partnership aimed at developing the computing foundation for utility-scale, fault-tolerant quantum systems. The collaboration brings AMD’s classical computing and software capabilities together with PsiQuantum’s photonic quantum computing architecture. AMD also intends to make an equity investment in PsiQuantum, although the companies have not disclosed its proposed value.
The announcement is significant because it demonstrates how quantum computing is attracting established technology suppliers whose expertise lies outside the development of quantum processors themselves. It also reflects an increasingly important reality: the practical capabilities of future quantum computers will depend on how effectively conventional computing infrastructure can support the specialised operations taking place within the quantum hardware.
Why quantum computers need classical partners
Quantum computing is often presented as a future alternative to conventional computing, but that description can obscure the relationship between the two technologies. The most credible development paths increasingly involve quantum processors operating alongside substantial classical computing resources.
A quantum processor performs specialised operations that exploit quantum behaviour. Operating that processor, however, requires conventional systems to prepare tasks, control hardware, process measurement results, detect and correct errors, and manage the flow of information through the wider computing environment.
Some of these activities must take place under demanding timing constraints. Quantum error correction, in particular, requires repeated measurements and rapid processing of information about potential errors. As systems become larger and more capable, the computational demands placed on the supporting infrastructure can increase considerably.
The objective is therefore not simply to attach a quantum accelerator to an existing computer. It is to develop an integrated architecture in which quantum and classical technologies perform the tasks for which they are best suited, with sufficient coordination to maintain reliable operation.
That is the opportunity AMD and PsiQuantum are seeking to address. Their collaboration illustrates how the development of quantum computing is becoming increasingly dependent on expertise accumulated within the conventional computing industry.
A partnership across the computing stack
The expanded relationship builds on existing work involving AMD’s Versal adaptive computing technology, which PsiQuantum selected for its fault-tolerant quantum computing platform.
The new agreement broadens the collaboration to include AMD EPYC processors, Instinct GPUs and the ROCm software platform. These technologies serve different functions within the classical computing environment, ranging from general-purpose processing to highly parallel computation and specialised low-latency tasks.
PsiQuantum contributes its silicon-photonics platform and the broader architecture it is developing for fault-tolerant quantum computation. The companies intend to co-design infrastructure capable of supporting large-scale quantum operations while also exploring applications that combine quantum computing with high-performance computing and AI.
The collaboration is consequently broader than a conventional hardware supply agreement. It reflects an attempt to bring together technologies that must eventually operate as parts of a complete computing system.
This matters because integrating components developed independently can introduce substantial engineering challenges. Performance, latency, communication interfaces, reliability and software compatibility all affect the behaviour of the overall platform. A highly capable quantum processor will not deliver its intended performance if the surrounding infrastructure cannot control its operations or process information quickly enough to maintain reliability.
The development of quantum computing is therefore increasingly about optimising entire systems rather than maximising the performance of individual components in isolation.
Error correction becomes an infrastructure problem
One of the most interesting elements of the partnership concerns quantum error correction.
Quantum information is unusually sensitive to errors caused by imperfect operations, noise and interactions with the environment. Fault-tolerant computing attempts to overcome those limitations by encoding quantum information in ways that allow errors to be detected and corrected.
The process requires extensive classical computation because measurement information must be interpreted so that the system can determine which corrections are necessary. These calculations become particularly demanding when they must be performed repeatedly and rapidly across large quantum systems.
AMD and PsiQuantum have already collaborated on PRISM, a machine-learning-based quantum error-correction decoder designed for implementation on programmable hardware accelerators. The companies say the work has been optimised to combine speed and accuracy for the low-latency requirements of PsiQuantum’s proposed architecture.
The significance lies in the relationship between quantum hardware performance and the classical infrastructure needed to sustain it. Improving the quality of quantum operations is essential, but the supporting system must also process the information required to keep those operations reliable. A powerful quantum processor connected to inadequate error-decoding infrastructure would face a fundamental operational constraint.
This is one reason why quantum hardware roadmaps increasingly involve conventional computing technologies. Advances in classical processors, specialised accelerators and software can influence the resources required to operate fault-tolerant quantum systems, potentially affecting their size, complexity and economic feasibility.
The eventual performance of a complete quantum computer will depend on how effectively these components operate together. Error correction is becoming as much a problem of systems architecture as of quantum physics.
Photonics and the manufacturing question
PsiQuantum’s approach also raises wider questions about industrial manufacturing. The company is developing quantum computing using photons rather than the superconducting or trapped-ion architectures pursued by several other major hardware developers.
Its strategy relies on silicon photonics and the potential to use established semiconductor manufacturing techniques to fabricate important elements of its technology. That connection to conventional semiconductor production is commercially attractive because existing manufacturing expertise and infrastructure could potentially support the development of components for larger quantum systems.
The relationship should not, however, be oversimplified. Using silicon photonics does not mean that manufacturing a utility-scale photonic quantum computer is equivalent to producing conventional semiconductor devices.
The complete system still requires specialised optical components, highly reliable networking, photon detection, control infrastructure and technologies capable of operating under demanding physical conditions. Some elements, including certain photon detectors, require cryogenic operation.
Photonic approaches therefore do not eliminate the infrastructure challenges associated with quantum computing; they change their character. Their commercial significance will depend on whether those requirements can be addressed at sufficient scale, reliability and cost.
This makes partnerships with established semiconductor and computing companies particularly relevant. Quantum hardware developers are increasingly dependent on manufacturing capabilities, engineering processes and supply-chain relationships accumulated through decades of conventional technology development.
A successful quantum architecture must ultimately be more than scientifically credible. It must also be manufacturable, supportable and capable of operating within a commercially viable infrastructure.
AI, quantum and high-performance computing converge
The AMD–PsiQuantum relationship illustrates another emerging pattern: quantum computing, artificial intelligence and high-performance computing are becoming more closely connected through their development methods and infrastructure requirements.
Classical supercomputers are already used to simulate quantum systems, investigate algorithms and evaluate potential architectures. Machine-learning methods can contribute to error correction and optimisation, while quantum computing researchers continue to explore applications where quantum and classical processing may eventually complement one another.
The expanded partnership includes work on computational chemistry, particularly drug discovery. According to the companies, PsiQuantum’s computational methods running on AMD hardware have produced encouraging simulation results involving covalent inhibitors, a class of molecules relevant to pharmaceutical research.
These findings need to be interpreted carefully. The reported results arise from classical computing methods developed through the collaboration; they do not demonstrate that a fault-tolerant quantum computer has already delivered a commercial advantage.
The companies are using available classical computing technology to improve computational approaches today while preparing for potential quantum capabilities in the future. That is an interesting strategy because it offers a route towards developing applications, expertise and commercial relationships before the arrival of large-scale fault-tolerant quantum systems.
Whether this approach eventually produces a sustainable market for quantum computing will depend on the value delivered and the economics of the resulting services. Nevertheless, it illustrates how the quantum industry is becoming integrated with established computing markets rather than developing entirely separately from them.
Building the industry before the market
The expanded partnership connects directly to the findings of State of Quantum 2026: Industry Ahead. Market Behind., published by The Quantum Space in October.
The report examines the development of quantum computing across hardware manufacturing, international supply chains, infrastructure, investment and commercial adoption. Its central argument is that the industrial capabilities required to support quantum computing are developing faster than the customer market that will ultimately need to sustain them.
The AMD–PsiQuantum announcement reinforces that assessment. Quantum companies are no longer operating solely within specialist research environments. They are establishing relationships with major technology suppliers, integrating conventional computing platforms and developing increasingly sophisticated industrial ecosystems.
These relationships can reduce technical uncertainty and strengthen the infrastructure needed for future quantum systems. They may also help establish clearer routes towards manufacturing, deployment and ongoing operational support.
The commercial market remains a separate question, however. An investment in quantum computing infrastructure is evidence of confidence in future demand, not evidence that the demand has already materialised. Similarly, an engineering partnership may make a future quantum computer more feasible without demonstrating that enterprises will find its services economically attractive.
This distinction is particularly important when discussing utility-scale quantum computing. The term suggests systems capable of delivering significant practical computing capabilities, yet their commercial usefulness will ultimately depend on which problems they can solve, how reliably they operate and how their costs compare with available alternatives.
Progress towards the ability to manufacture and operate more advanced machines is essential. It does not automatically establish the value proposition for customers expected to purchase their services.
The supplier ecosystem takes shape
There is also a strategic dimension to the partnership. As quantum systems become more complex, their development will depend increasingly on international networks of specialist suppliers and manufacturing organisations.
Advanced semiconductors, optical components, computing accelerators, cryogenic equipment, networking technologies and software platforms may all contribute to the operation of future machines. That creates opportunities for companies whose principal markets are not currently quantum computing, while also introducing dependencies that will need to be understood and managed.
A quantum hardware developer may be capable of designing a sophisticated computing architecture while remaining dependent on external suppliers for essential components. Some of those technologies may be highly specialised, difficult to replace or concentrated within a small number of manufacturers.
The ability to secure reliable industrial partnerships will therefore influence which quantum architectures can be developed and deployed commercially. Supplier relationships may become as strategically significant as intellectual property within individual quantum processor designs.
For governments pursuing quantum sovereignty strategies, this is an important consideration. Domestic expertise in quantum processor development does not necessarily provide control over the wider technologies needed to construct a complete system.
A credible industrial strategy needs to understand the full supply chain, including the classical infrastructure supporting the quantum hardware. The AMD partnership illustrates how deeply these industrial ecosystems are beginning to overlap, and why national quantum ambitions must be considered within the wider context of advanced computing and semiconductor manufacturing.
What would constitute commercial progress?
For investors and enterprise customers, the relevant measures of progress are becoming more demanding. A partnership announcement provides useful evidence of industrial development, but it is not sufficient evidence of commercial maturity.
Future assessments will need to consider whether the collaboration produces functioning infrastructure, measurable performance improvements and systems capable of supporting useful applications. In error correction, that means evaluating whether the classical processing architecture can support the requirements of larger fault-tolerant machines. In application development, it means establishing whether quantum-classical workflows offer meaningful advantages over existing computing methods.
For customers, the ultimate assessment will involve reliability, accessibility, operating costs and integration into established business processes. These questions cannot be answered through investment announcements alone; they require demonstrations, deployments and evidence of repeatable customer value.
The quantum industry is becoming increasingly capable of explaining how its future machines might be built. The next challenge is demonstrating why organisations will need to use them and under what commercial conditions those services will become attractive.
Our companion Innovating Trust episode, Quantum’s Trillion-Dollar Question, explores precisely this relationship between technological development, industrial investment and the emergence of a sustainable commercial market.
A different kind of quantum milestone
The AMD–PsiQuantum announcement is not a record-breaking qubit demonstration, nor is it proof that commercially useful fault-tolerant quantum computing has arrived. Its significance lies in providing another indication that quantum computing is evolving towards an industrial discipline in which the performance of the complete system matters more than any individual technological component.
Classical processors, specialised accelerators, networking, error correction and manufacturing infrastructure are becoming increasingly important parts of the engineering equation. Established computing companies have valuable expertise to contribute, and the partnerships being formed today may influence which quantum technologies can be built and supported at scale.
This represents genuine industrial progress, but it also reinforces the question raised in State of Quantum 2026. The industry continues to assemble the technology, infrastructure and partnerships required for the next phase of quantum computing, while the commercial market must develop sufficiently to justify that investment.
Building a quantum computer is becoming an increasingly complex industrial undertaking. Building a market for quantum computing remains a separate commercial challenge, and the relationship between those two developments will determine how quickly the technology moves from engineering ambition towards sustainable enterprise adoption.




Leave a Reply