The field of quantum computing is filled with lofty goals, but few have committed to scale as aggressively as PsiQuantum. While most companies chase incremental gains, like more stable qubits, longer coherence times, and smaller error rates, PsiQuantum has declared that it is building a machine with one million qubits. That system is not a distant aspiration. Erik Hosler, a panelist attentive to the relationship between manufacturing scale and quantum system design, noted that the effort to produce it is already happening. It is not a proposal, but a production roadmap.
What makes PsiQuantum’s approach remarkable is not just the number. It is a fact that the company has placed fabrication and manufacturability at the center of its architecture. From the beginning, the platform was designed to be compatible with advanced photonics fabrication, wafer-scale integration, and global semiconductor infrastructure. The million-qubit target is not an abstraction. It is a function of their chosen tools.
A Million Qubits Is Not a Theory
Most quantum hardware developers are cautious with their projections. They move from dozens of qubits to hundreds, then plan for low thousands. PsiQuantum has gone in the opposite direction. It has announced an intent to skip directly to the scale that some experts believe is required for practical fault-tolerant computation.
It is not a branding exercise. It is a deeply considered bet. The company’s leadership believes that smaller systems, while valuable for research and proof-of-concept demonstrations, cannot have a commercial impact. Only a machine with hundreds of thousands of logical qubits and built-in error corrections can unlock the kinds of workloads quantum computing is expected to manage.
This belief shapes their engineering strategy. PsiQuantum does not wait for every piece of the system to be perfected before scaling. Instead, it is designed for scale from the beginning, accepting that some subsystems will need to improve in parallel. They assume that manufacturing is not the last step. It is a co-equal pillar of system design. That strategy would be reckless if the manufacturing tools were not ready. But PsiQuantum is betting that they are.
From Vision to Fab Floor
That bet was underscored during the SPIE Advanced Lithography panel. “PsiQuantum aims to build a million-qubit system, with manufacturing already underway,” Erik Hosler observes. This statement carries real weight. “Already underway” in this context does not mean a few test runs or custom wafers. It means full-scale planning for packaging, layout, photonic routing, and integration with cryogenic systems. It signals that PsiQuantum has moved from cleanroom research to production partnerships. The company has announced its use of GlobalFoundries to fabricate parts of its quantum architecture using photonic processes. That decision reflects a shift from R&D to yield-aware fabrication.
It also shows that quantum computing, once confined to physics departments and national labs, is stepping into the workflows of commercial chip manufacturing. Tools once developed to print logic transistors are being reused to build entangled photon sources and beam splitters. Mask aligners, resist chemistries, and optical metrology tools are all part of the effort to scale. The significance lies not in the quantity of qubits alone, but in the transition from lab concept to fabrication reality.
Built Like Chips, Not Lab Experiments
PsiQuantum’s architecture relies on photonics, and that choice is not incidental. Photonic qubits have several advantages. They operate at room temperature. They can be manipulated using known optical components. Most important, they can be built on silicon wafers using mature foundry tools. It makes PsiQuantum’s approach fundamentally different. Its devices are not assembled manually on test benches. They are printed and etched on wafers. Photonic paths, interferometers, and detectors are patterned using the same systems used for classical circuits.
This strategy reflects a key theme from the SPIE panel. Patterning, etch, and metrology practices developed for CMOS are now being applied to quantum. PsiQuantum’s platform does not ask fabs to adopt new tools. It asks them to apply existing ones to a new objective. There are challenges. Photonic systems demand extremely precise control over alignment, path length, and interface quality. But the core lesson holds. By building quantum systems like chips, PsiQuantum opens a pathway to scale that sidesteps the limits of hand-crafted approaches.
The Risk and Reward of Going Big First
Scaling before full validation involves risk. Building a million-qubit machine without running a ten-thousand-qubit version first commits the team early. If foundational issues emerge, the impact is magnified. Some quantum developers take a different path. They add qubits gradually, learning with each step. It offers time to adjust and refine. But it also slows the move toward systems that can do useful work. It may yield better prototypes, but not always systems that scale efficiently.
PsiQuantum has accepted that risk. They are betting that starting large will surface challenges quickly and that early integration with the semiconductor ecosystem will provide the tools to solve them.
It is not an unusual mindset in classical chipmaking. Build early, fix fast, and iterate using data from real production environments. That philosophy is now being applied to quantum.
What a Million Qubits Really Demands
Reaching a million qubits is not just a technical exercise. It forces design teams to solve system-level problems that smaller platforms can defer. It requires answers for packaging density, control line routing, photon loss budgets, and real-time calibration.
These questions lead beyond physics and into engineering. They require tooling that supports repeatability, automation that maintains stability, and integration frameworks that minimize latency and signal degradation. PsiQuantum’s approach confronts these challenges directly. Rather than waiting for ideal conditions, the team has aligned design decisions with real manufacturing constraints. It makes the platform not just ambitious, but grounded.
Building this scale is difficult. It demands coordination across disciplines, from photonics to materials science to fabrication engineering. But it may be the only way to transition from experimental promise to real-world application. The path to quantum utility is not defined only by qubit count. It is defined by how those qubits are built, connected, and made to work together. And for PsiQuantum, that process has already begun.
