Researchers have made progress in the global race to build bigger and better quantum computers. A new machine called Helios is a fundamentally different kind of system from other quantum computers.
Quantum computers harness the power of quantum mechanics, which governs the laws of physics at the atomic and subatomic scale. Among the various designs for such machines, Helios is a trapped-ion quantum computer, which uses charged atoms suspended in free space by electromagnetic fields.
It operates with 98 qubits, the units of information that quantum computers use to process data. That makes it the largest trapped-ion quantum computer ever built. Quantinuum, the company that developed the device, is based in Cambridge, UK, and Broomfield, Colorado, in the US. It had previously demonstrated earlier machines operating with 32 qubits in 2023 and 56 qubits in 2025.
Helios and its predecessors use an architecture (or operating structure) with separate regions for storing and processing quantum information. This architecture, known as QCCD (quantum charge-coupled device), was proposed in 2002. It resembles the architecture of classical computers, which have memory for storage (hard disk drives, solid-state drives) and separate processors (CPUs, GPUs).
The specific geometry of Helios resembles a rosette, consisting of a ring for storage and two streamers for processing. The key element is a four-way X junction where these meet. QCCD physically transports charged particles (ions) electrically from the storage region to the processing regions.
Processing is carried out using laser pulses. This requires quantum algorithms to be broken down into ion transport and quantum processing. This differs significantly from other quantum computing architectures, such as superconducting qubits, in which the qubits are spatially fixed and processing is carried out by electrical signals that are switched on and off in time.
Error correction
Classical charge-coupled devices, which were the core component of early digital cameras, are the prototype for the approach now used in Helios. Using this approach for quantum computation has three main advantages.
Gates are the basic building blocks of quantum circuits, and they perform operations on qubits. One advantage of the QCCD approach is that an operation on a particular qubit does not affect neighboring qubits. This reduces a phenomenon known as "crosstalk," leading to higher-quality performance.
The second advantage is that qubits can be measured and reset mid-computation. This allows errors, which are inevitable during computation, to be detected and corrected as early as possible, improving performance further.
Finally, after processing, qubits can be (within certain constraints) called up from, or returned to, widely separated parts of the memory. This makes it possible to connect distant qubits, enabling more efficient operation.
Helios has demonstrated all of these advantages. This is made possible by two major advances, in hardware and software respectively. First, the four-way X junction allows the system to handle multiple tasks at once rather than one at a time, leading to a significant speedup. This was not possible with earlier QCCD machines, which could only move data along a single line or loop.
Second, a new classical control software called Helios runtime was developed to make clever use of the two-dimensional freedom the X junction provides. It plans the most efficient and fastest routes for moving and processing data.
Together, these represent a major advance in the engineering of quantum computers. An important consequence is that Helios can perform calculations that, using known methods, even the largest supercomputers could not run within a realistic time and power budget.
Computational power
This suggests that the device and its successors have capabilities that surpass the computational power of classical computers, although classical algorithms and hardware are also continuing to advance rapidly.
Nevertheless, Helios's computations so far are merely random benchmark tests, and the practical significance of this leap remains limited. To carry out quantum computations of practical importance in science and commerce, quantum computers will need to be orders of magnitude larger and more capable than they are now, on the order of roughly a million qubits, according to even the most optimistic estimates.
Mission 1 of the UK National Quantum Strategy sets a goal of realizing a fault-tolerant quantum computer (FTQC) capable of running a trillion operations. For reference, the longest computation Helios has run was about 4,000 operations on 98 qubits.
Building an FTQC from Helios, or from any other known architecture or system, remains a major challenge. As these numbers suggest, it would likely require connecting thousands of QCCD devices, each with thousands of qubits, via quantum links.
There are proposals for QCCD devices with a two-dimensional grid structure, like a city's street network, that could operate with thousands or more qubits. However, the unavoidable challenge of engineering and operating such systems, with huge numbers of junctions and the resulting qubit "traffic jams," a phenomenon akin to road congestion, has not yet been faced, let alone overcome. The physical transport of ions is also a very slow process, a consequence of the need for extremely precise voltage control. This problem is expected to become even more severe as devices grow larger and qubit counts rise.
For now, the Helios device is a landmark advance in the field of quantum computing. Whatever form future FTQCs take, they will require both creatively designed quantum hardware devices and the sophisticated software that supports them.