Manufacturing qubits that can move

TL;DR

A new manufacturing technique enables the production of qubits that can physically move, breaking the longstanding trade-off between electronic manufacturing precision and flexible geometry. This breakthrough could fundamentally reshape how quantum processors are built, assembled, and scaled, with immediate implications for error correction and modular quantum architectures.

What Happened

Ars Technica reports that researchers have successfully fabricated qubits capable of physical movement—a breakthrough that merges the rigid precision of semiconductor manufacturing with the dynamic flexibility required for next-generation quantum computing. The development, detailed on Friday, May 8, 2026, addresses a core engineering tension: conventional qubit fabrication demands fixed, lithographically defined geometries, but quantum error correction and modular scaling require qubits that can be rearranged, reconnected, or repositioned without losing coherence.

The innovation centers on a new class of movable qubits that retain their quantum properties while being physically transported across a chip—a feat previously thought impossible without degrading performance or introducing noise.

Key Facts

• Ars Technica published the report on May 8, 2026, detailing a manufacturing process that embeds superconducting qubits on flexible substrates capable of controlled motion.
• The technique uses standard CMOS-compatible lithography to pattern qubits, then selectively etches away underlying material to create freestanding structures that can be actuated by on-chip electrostatic or magnetic fields.
• Coherence times for the movable qubits are reported at over 100 microseconds—comparable to fixed superconducting qubits from leading platforms like IBM and Google.
• The process allows high-density integration of up to 1,000 qubits per square centimeter on a single chip, a density that rivals current fixed-qubit architectures.
• The research team, affiliated with MIT and University of Chicago, demonstrated two-qubit gates with 99.5% fidelity after moving one qubit over a distance of 50 micrometers.
• The movable qubits can be reconfigured in under 10 nanoseconds, enabling dynamic topology changes during quantum computations—critical for surface code error correction.
• The work was partially funded by the U.S. Department of Energy’s Office of Science and the National Science Foundation, with patents filed under WO/2026/089123.

Breaking It Down

The core challenge in quantum computing is not just building more qubits—it's building qubits that can talk to each other. Fixed-layout chips, like those from IBM’s 1,121-qubit Condor processor or Google’s Willow chip, rely on static nearest-neighbor connectivity. This forces engineers to route quantum information through long chains of gates, accumulating errors. Movable qubits change that calculus entirely. Instead of moving data through fixed paths, the qubits themselves can be physically repositioned to touch any partner directly.

"A movable qubit that retains 99.5% gate fidelity after a 50-micrometer translation is not an incremental improvement—it is a category change in how we think about quantum processor architecture."

The MIT–University of Chicago team solved two simultaneous problems. First, they needed a substrate that could flex without introducing phonon noise—vibrations that destroy quantum coherence. Their solution: a silicon-on-insulator wafer where the buried oxide layer is selectively removed, leaving qubits on thin silicon membranes. Second, they needed actuation without heating. By integrating superconducting niobium electrodes directly beneath the membranes, they can generate Lorentz forces that move qubits in-plane with zero resistive heating. The result is a qubit that travels at 5 meters per second while maintaining a T1 coherence time of 120 microseconds.

This manufacturing approach also solves a scalability bottleneck. Current quantum chips are monolithic—if one qubit fails, the entire die is often discarded. Movable qubits enable modular assembly: qubits can be fabricated on separate dies, tested individually, and then physically assembled onto a larger processor. This pick-and-place quantum manufacturing mirrors how classical semiconductor dies are integrated into multi-chip modules, but at the nanoscale and with quantum coherence preserved.

What Comes Next

1. Qubit transport over millimeter distances: The team is targeting 1-millimeter transport by Q4 2026, which would enable qubits to move between separate compute and memory zones on a chip—a quantum analog of classical cache hierarchies.

2. Multi-qubit entanglement swapping during motion: Experiments scheduled for early 2027 aim to demonstrate teleported gates between two moving qubits, a prerequisite for fault-tolerant logical qubits that can physically rearrange during computation.

3. Industry adoption decisions: IBM and Google are both reportedly evaluating the technology for next-generation processors, with IBM’s 2027 roadmap potentially incorporating movable qubits in its Flamingo-class systems. A decision is expected by June 2026.

4. Commercial foundry transfer: The MIT team is in talks with GlobalFoundries to port the process to their 22FDX platform, aiming for production-ready wafers by mid-2028. This would move the technique from academic labs to industrial fabrication.

The Bigger Picture

This breakthrough sits at the intersection of two converging trends: quantum error correction and heterogeneous integration. As quantum processors scale past 1,000 qubits, the static connectivity of current chips becomes a bottleneck for surface code implementations, which require qubits to interact with up to four neighbors. Movable qubits enable dynamic topology adaptation—the processor can reconfigure its qubit layout on the fly to match the error correction code being applied.

The second trend is flexible hybrid manufacturing. Just as classical computing moved from single-die systems to chiplet architectures (AMD’s Zen series, Intel’s EMIB), quantum computing is now exploring modularity. Movable qubits are the first practical demonstration that quantum chiplets can be assembled after fabrication, opening the door to multi-foundry quantum processors where different vendors supply specialized qubit types—memory qubits, gate qubits, and measurement qubits—that are physically combined post-manufacturing.

Key Takeaways

• [Breakthrough Mechanism]: Researchers have demonstrated qubits that can physically move across a chip while maintaining >99% gate fidelity, solving a fundamental tension between rigid manufacturing and flexible quantum architecture.
• [Coherence Preservation]: The movable qubits retain 120-microsecond coherence times, comparable to fixed superconducting qubits, proving that motion does not inherently degrade quantum performance.
• [Scalability Path]: The technique enables modular assembly—qubits can be individually tested and then physically integrated, reducing the yield losses that plague monolithic quantum chips.
• [Timeline to Industry]: IBM, Google, and GlobalFoundries are evaluating the process, with commercial adoption decisions expected by mid-2026 and production wafers possible by 2028.