147
Part Three · E · Hardware, simulation and practice

Photonic and measurement-based computing

Photons are ideal messengers: they travel at the speed of light, barely interact with their surroundings, and keep their quantum state across kilometres of fibre at room temperature. That same aloofness makes them terrible at logic, because a two-qubit gate needs two photons to interact. Every photonic quantum computing scheme is a workaround, and the most powerful one changes what computing means: prepare a large entangled resource first, then compute purely by measuring it.

On the plate: a computation driven only by measurements.

KLM: gates from measurement

Emanuel Knill, Raymond Laflamme and Gerard Milburn showed in 2001 that linear optics (beam splitters, phase shifters), single-photon sources and detectors suffice for universal quantum computing. The interaction is replaced by measurement: the Hong–Ou–Mandel effect (Chapter 43) and post-selection make a gate that succeeds only some of the time, heralded by which detectors click. Raising the success probability needs many extra photons, so KLM in its original form is extremely costly. It proved possibility, and pointed to a better idea. Established

Measurement-based computing

Robert Raussendorf and Hans Briegel showed in 2001 that a cluster state, qubits prepared in ∣+⟩|+\rangle and entangled by CZ gates between neighbours, is a universal resource. Measuring one qubit in the basis (∣0⟩±eiφ∣1⟩)/2(|0\rangle\pm e^{i\varphi}|1\rangle)/\sqrt2 teleports its state onto the next qubit with the gate H diag(1,e−iφ)H\,\mathrm{diag}(1, e^{-i\varphi}) applied, up to a known Pauli correction depending on the random outcome. Choose an input state angle 0.80, measurement angles φ1=\varphi_1 = 0.50, φ2=\varphi_2 = 1.20, and . The plate simulates the entangled chain, draws random outcomes, applies the corrections and compares the output with the intended gates: fidelity 1 whatever the outcomes. The blind-computing protocol of Chapter 131 is built on exactly this. Established

Fusion, loss, and the companies

Large cluster states are built by fusing small entangled photon states with two-photon measurements that succeed 50% of the time without extra resources. Sara Bartolucci and colleagues at PsiQuantum formalized fusion-based quantum computing in 2023, designing codes that tolerate fusion failures and photon loss. Loss is the enemy: fibre loses about 0.2 dB per kilometre, so over 100 km the surviving fraction is 10−0.02L10^{-0.02L}. In January 2025 Xanadu reported Aurora, a modular machine of 35 photonic chips linked by 13 km of fibre, using squeezed light and GKP-type encodings (Chapter 141). PsiQuantum described a manufacturable chipset, Omega, in Nature in 2025; its claims about manufacturing scale come largely from the company. Established for the physics; Debated for the timelines.

What you now have

Why photons resist logic, the KLM scheme, cluster states and measurement-based computing with byproduct corrections, fusion-based computing, loss budgets, and the 2025 photonic systems.

· · ·

Three questions to carry with you.

What replaces interaction in linear-optics computing?tap
Measurement and post-selection, heralded by detector clicks.
What does measuring a cluster qubit at angle φ do?tap
Teleports the state to the next qubit with H·diag(1, e^(−iφ)) applied, up to a known Pauli correction.
Why is loss the critical photonic error?tap
A lost photon removes a qubit entirely; codes and fusion networks must be designed to tolerate it.

And one to work out. At 0.2 dB/km, what fraction of photons survives 100 km of fibre?

fraction =

Sources: E. Knill, R. Laflamme and G. J. Milburn, Nature 409, 46 (2001). R. Raussendorf and H. J. Briegel, Phys. Rev. Lett. 86, 5188 (2001). S. Bartolucci et al., Nat. Commun. 14, 912 (2023). H. Aghaee Rad et al., Nature (January 2025) (Aurora). PsiQuantum team, Nature (2025) (Omega). Last checked 18 November 2026. Checked by the plate: exact two-qubit simulation of each measurement step, fidelity 1 for every outcome pattern.

Next · Chapter 148
Spin and topological qubits

One company says it has built a topological qubit. Many physicists aren't convinced. Here is how to read the evidence yourself.

Turn the page