Google Patents a Tunable Switch That Connects Quantum Bits With Less Hardware
Building a quantum computer isn't just about making better qubits, it's about connecting them cleanly without burying the whole system in control wires. Google's latest patent tackles exactly that problem.
What Google's qubit coupler actually does
Every time an engineer runs a calculation on a quantum computer, dozens of qubits (the quantum equivalent of a regular computer's bits) have to talk to each other at precisely the right moment and go silent just as precisely afterward. Getting that timing right, across many qubits at once, requires a lot of control hardware, and that hardware is one of the biggest obstacles to scaling quantum machines up.
Google's patent describes a smarter coupling circuit that sits between pairs of qubits. By adjusting the energy levels inside that circuit, a controller can turn the connection between two qubits on or off without needing separate, complex wiring for each pair. The circuit has at least two internal modes, and the controller uses those modes like dials to dial the coupling strength up or down.
The specific qubits involved are called fluxonium qubits, a design that tends to hold their quantum state longer than older designs. Combined with this tunable coupler, Google is trying to make a system where you get both better qubit quality and a simpler control setup at the same time.
… adjust one or more energy levels of the at least two modes of the coupling circuit to thereby increase or decrease a coupling strength between the first fluxonium qubit and the second fluxonium qubit via one or more of the at least two modes of the coupling circuit.
Translation: It changes the internal energy levels of the connector to dial the connection between two qubits up or down.
How the coupling circuit controls qubit interactions
At the center of this patent is a coupling circuit that connects two fluxonium qubits. Fluxonium qubits are a relatively newer design that stores quantum information in a way that makes them less sensitive to certain types of noise, which means they tend to hold their state (called coherence) longer than older transmon-style qubits.
The coupling circuit has at least two internal modes (think of modes as distinct ways the circuit can oscillate, each carrying a different energy). A controller adjusts the energy levels of those modes, which changes how strongly each mode interacts with the neighboring qubits. Raising or lowering that energy is like moving a bridge closer to or further from two islands, it controls whether the islands can exchange traffic.
This design enables two key things:
- Tunable coupling strength: the controller can increase or decrease how tightly the two qubits are linked, including switching the link fully off when the qubits need to stay isolated.
- Entangling gates: by driving the coupler or the qubits themselves between their energy states, the system can perform entangling operations (the two-qubit logic steps that give quantum computers their power) without requiring as many separate control lines.
The goal is a cleaner, more scalable architecture: fewer physical control components per qubit pair, which matters enormously when you want to build a chip with hundreds or thousands of qubits.
Superconducting qubits may be coupled together via a tunable coupler that may be controlled to adjust the energy of one or more states of the coupler, allowing the extent to which states of the coupler couple to states of the qubits to be adjusted.
Translation: A flexible connector links the qubits by letting engineers tune its energy states to control how they interact.
What this means for building bigger quantum chips
Scaling quantum computers is less about raw qubit count and more about control overhead. Every additional qubit today typically requires additional classical electronics to drive and read it. If that overhead grows faster than the qubit count, you hit a ceiling, and current quantum hardware is bumping against it.
A tunable coupler that reduces the number of control lines per qubit pair is a direct attack on that ceiling. Google's run of quantum-hardware filings suggests the company sees architectural simplification as the near-term path forward, not just raw qubit improvements. For anyone following the quantum computing race, this kind of infrastructure work is where the real competition is happening right now, even if it's far less visible than headline qubit numbers.
This is the eighth Google filing we've tracked in quantum computing since May, adding to work like one fixing mid-cycle errors and one grading gates one by one.
Google's approach here is pure hardware, meaning it cannot be delivered as a software update the way a phone feature can. The design calls for new physical circuits that must be built, cooled to extreme temperatures, and tested before anything else happens.
Getting from a promising circuit design to a reliable product requires solving manufacturing problems that compound as you add more qubits to the system. Each step of that process, design, fabrication, testing, and redesign, typically takes years.
What the patent does show is that Google is attacking the right bottleneck. Simplifying the wiring and control equipment needed to coordinate qubits is exactly what has to happen before a quantum computer can scale from a research curiosity to something practically useful, and the choices made at this architectural stage will determine what becomes possible down the road.
There are more where this came from
We read every patent application Big Tech publishes and send you the ones worth knowing. Plain English, free, every week.
The drawings
26 drawing sheets from US 2026/0260146 A1 · click any drawing to enlarge
Want this weekly breakdown for a company we don't cover? Patentlyze Pro →