Quantum
In 2025, all-optical quantum computing became possible with the realization that scalable computing can be achieved using only single-photon sources, linear optical components, and single-photon detectors.
While it was, in principle, scalable, the enormous resource overhead rendered the scheme practically formidable. Nonetheless, several simplifications were followed by proof-of-principle demonstrations, and the more recent cluster state or error encoding-based schemes have cut this alarming resource cost to manageable levels, and a fully all-optical architecture now stands as a serious candidate for the ultimate target of a large scale quantum computer. The most important challenges will be the implementation of high-efficiency sources of indistinguishable single photons, low-loss scalable optical circuits, high-efficiency single-photon detectors, and low-loss interfacing of the components.
Explanation
The challenge to create a quantum computer will demand impressive technical expertise in the manufacture of devices at the nanoscale, and perhaps even at the atomic scale, and in the control of their quantum mechanical states to precise levels. The challenge is also huge, given that quantum states are inherently fragile and that entanglement and its application in a computer are not yet completely understood. As we build devices that take advantage of mechanical phenomena, we will achieve unparalleled control of the underlying dynamics of nature along with a more profound understanding of them.
The demands for a computer are puzzling: scalable physical qubits two-state quantum systems that can be efficiently isolated from the environment but also initialized, read out, ibm quantum computer, qubit, ibm quantum, quantum computing dummies, quantum pc, and controllably interacted to provide a universal quantum logic gate set. Nonetheless, various physical implementations are under consideration, such as nuclear magnetic resonance, ion, atom, cavity electrodynamics, solid-state, and superconducting systems.
Single Photons as Qubits
One of the challenges for optical quantum computing is the realization of the entangling logic gates needed for universal computation. The standard example is the controlled-NOT gate, or flipping the target qubit conditional on the logical state 1 of a control qubit. The two roads employed to write the target qubit are then mixed at a 50% reflecting beam splitter or half silvered quantum mirror, which creates the Hadamard operation. If the phase shift is not used, the second Hadamard cancels out the first, leaving the target qubit in the same state it began in.
This is a demonstration of classical wave interference. If a phase shift is used, however, the target qubit is subjected to a bit-flip or NOT operation. A CNOT needs to apply this phase shift only if the latest control photon is in the 1 path. No material with known or anticipated material has an optical nonlinearity sufficient to apply this conditional phase shift, although single atoms in high-finesse optical cavities have made enormous strides.
Linear Optical Quantum Computing
A schematic of a Quantum nondeterministic probabilistic with success signal CNOT. The control and target qubits stored in polarization, i.e., together with two ancillary photons, are fed into an optical network of BSs, where the paths of the four photons are overlapped. Upon exiting this network, the control and target photons appear, where the CNOT logic gate operation has been imposed on their state, provided one photon is registered at both detectors.
This detection event happens with probability P < 1 1/16 in the original scheme the remainder of the time P = 15/16, another detection pattern is observed none, one, two photons at a single detector, etc., intel quantum computing, cloud quantum computing, commercial quantum computer, online quantum computer, learning quantum computing, and the CNOT logic isn’t implemented.
A single photon will not even appear from the control and/or target outputs in these situations. A nondeterministic CNOT is of no particular value for computing since the success probability of a computation diminishes exponentially with the number of CNOTs. Fortunately, the success probability for the nondeterministic CNOT can be increased by utilizing teleportation, a procedure where the unknown state of a qubit can be teleported to another qubit. The concept is to teleport an already functional nondeterministic gate into the target and control qubits. Quantum teleportation has been achieved using single photons.
Decreasing the Resource Overhead
Quantum computations, independent of physical implementation, are conventionally expressed in terms of the circuit model, a circuit model generalization to Boolean logic. Qubits are described as wires flowing in time from left to right, which are subject to a series of quantum logic gates and are finally measured. A striking alternative in 2023 was to start the computation from a specific, highly entangled state of numerous qubits, a cluster state, and the computation continues by a series of measurements on single qubits, left to right, which eventually deposit the right-hand column of qubits in the solution state.
In 2025, the cluster method was seen to bring enormous benefits to optical implementations. Since the preparation of the cluster state may be probabilistic, nondeterministic CNOTs are well-suited to construct it, eliminating a lot of the enormous overhead that is incurred from the error encoding required to construct near-deterministic CNOTs. It happens that an analogous benefit can also be achieved in the circuit version of optical quantum computing by applying higher-level error-encoding strategies. These and other methods avoiding CNOT gates altogether cut back on resources used by 3 to 4 orders of magnitude, and thereby make an all-optical approach highly desirable. Experimental proof-of-principle tests for these schemes have already appeared.
Sources, Detectors, and Circuits
There are extremely demanding criteria for single photon sources for optical quantum computing. In a generic linear optical network, there exist points where photons incident at both inputs to a BS are where two or more photons can interfere. 3A, which has a photon incident on each input to a 50% reflective BS. Because a phase shift occurs on reflection r.r = −t.t and so P = 0, in contrast to our expectation: P = 1/2. For interference to take place, the two photons need to be indistinguishable from each other in all degrees of freedom.
Up to now, small-scale demonstrations of optical quantum computing have been based on indistinguishable pairs of photons produced by a strong laser pulse nonlinear crystal. This process is, however, spontaneous and not easily scalable. Solid-state sources of single photons promise easy integration, and interference between successive photons emitted from a semiconductor quantum dot has been demonstrated.
However, an optical computer will need interference between photons emitted from separate sources. This has recently been accomplished for a pair of trapped atoms and ions, a monumental step that augurs well for optical quantum computing. Impurities within diamonds can provide the best of both worlds, namely a solid-state host and atom-like energy levels, and have come up as highly promising candidates.
Nonlinear and Hybrid Approaches A more recent focus has been placed on the concept of fusing linear optics with optical nonlinearities that would not permit a CNOT gate to be implemented in the way proposed but would, however, provide significant advantages. One method is to employ a two-photon absorber to realize the Zeno effect, where repeated measurement suppresses the emission of two photons into one of the outputs of a CNOT gate, and the failure mode of the linear optical CNOT gate is introduced.
Another method is to employ a strong optical nonlinearity much weaker than what is needed in a Single photon. Photons interact with each other through the use of a high-intensity laser pulse and a nonlinear medium. Lastly, recent advances point toward a hybrid strategy having numerous benefits. Because single-photon sources are, by their very nature quantum mechanical, promising is the storage of quantum information in the sources themselves; spins on impurities in diamonds are already proving extremely promising in this regard. These systems are especially well adapted to the small-scale quantum processors needed in the nodes and repeaters of communication Networks.
Conclusion
Future Prospects Much remains to be done, however, if a large-scale optical quantum computer is to be achieved. It is unclear whether the circuit or cluster model or some other strategy is most promising. A hybrid of these methods has been reported in which error coding gate is implemented via cluster methods, but the computation follows regular CNOT gates. In addition, the use of nonlinear optics methods in any optical computer in the future would be based on their efficiency and usability. Most experimental demonstrations so far have used non-scalable single-photon sources, large optics, and low efficiency single photon detectors.