Variational Quantum Eigensolver
VQE estimates the lowest energy of a molecule or material by running a short parameterized circuit, measuring the energy of the state it prepares, and letting a classical optimizer adjust the circuit's parameters — a hybrid quantum–classical loop built for today's noisy hardware. It genuinely runs on real machines at small scale, but it has no proven speedup, and classical chemistry methods currently match or beat every published VQE result.
| Task | Best classical | Quantum |
|---|---|---|
| Exact ground-state energy | Exponential in system size (full configuration interaction) | Unknown — VQE is a heuristic with no runtime or accuracy guarantee |
| Where practice stands today | CCSD(T), DFT and DMRG handle systems far larger than any quantum demo, at chemical accuracy for many of them | ~12-qubit demos with error mitigation, accuracy below classical baselines |
What problem does VQE try to solve?
Finding the ground-state energy of a quantum system — the single most useful number in computational chemistry, because energies determine reaction rates, binding strengths and material properties. The state space of n electrons-in-orbitals grows exponentially, so exact classical computation caps out at small molecules.
The proper quantum algorithm for this is phase estimation, but its circuit depths need fault tolerance. VQE, proposed in 2014, is the workaround: what is the best we can do with circuits shallow enough to survive today's noise?
The classical approach
Classical quantum chemistry is a mature, formidable field. Density functional theory handles thousands of atoms approximately; coupled cluster — CCSD(T), the "gold standard" — reaches chemical accuracy for medium-sized molecules; DMRG and tensor networks tackle strongly correlated systems. These methods run the world's actual chemistry workloads.
The genuinely hard classical cases are strongly correlated systems — transition-metal catalysts such as the nitrogenase FeMo cofactor is the standard example. That is the target quantum methods aim at. Whether VQE specifically can ever beat classical methods there is an open question, not a promise.
The quantum approach: a guess-and-check loop
VQE rests on the variational principle: any trial state's energy is greater than or equal to the true ground-state energy, so lower is always better. The loop:
- A parameterized circuit (the ansatz) — rotation gates with adjustable angles plus entangling gates — prepares a trial state.
- Repeated runs and measurements estimate the energy ⟨H⟩ as a weighted sum of simple expectation values, one measurement basis per group of terms.
- A classical optimizer nudges the angles and repeats.
Note what the quantum computer contributes: it evaluates one candidate state per run. All the searching is classical. The hope is that an entangling ansatz can represent correlated electronic states more cheaply than classical descriptions — plausible in principle, unproven in practice.
Run it: one energy evaluation inside the loop
Limitations
- No guarantees. Unlike phase estimation, VQE promises nothing: the ansatz may be unable to represent the ground state, and the optimizer may never find the best parameters it can represent.
- Barren plateaus. For many ansatz families, energy gradients shrink exponentially with qubit count, leaving the optimizer wandering a flat landscape. This is a proven obstacle for broad ansatz classes, not a rumour.
- Measurement cost. Chemical accuracy (~1.6 milli-Hartree) demands millions to billions of shots per energy evaluation for realistic molecules — often the dominant, and quietly ignored, cost.
- Classical shadow. Shallow circuits that noise permits are frequently classically simulable, and recent classical surrogate methods reproduce many published VQE experiments outright — squeezing the window for advantage from both sides.
On hardware today
VQE is among the most-run algorithms on real hardware: H₂ on 2 qubits (2016), BeH₂ on 6 (2017), a 12-qubit Hartree–Fock calculation (2020), and many Hubbard-model and small-molecule experiments since, across superconducting and trapped-ion machines — see current QPUs. Every one required error mitigation, and every one computed energies a laptop gets faster and more accurately.
That is not damning — these are hardware benchmarks, honest as such. But no VQE calculation has ever exceeded classical capability, and the field's own trajectory has shifted toward error-corrected algorithms as the credible route to useful quantum chemistry. Run the ansatz in the lab, and note how transpilation reshapes even this 5-gate circuit on real devices.