Chemistry & materials
Molecules are quantum systems, so simulating them is a natural fit. Better catalysts, batteries and drugs are the long-term prize.
A superconducting quantum computer looks like a golden chandelier hanging in a steel can. Almost everything you see exists for one reason: to keep a chip the size of a coin cold, still and silent.
Empty space sits at about 2.7 kelvin. The bottom of this machine reaches roughly 15 millikelvin, well over a hundred times colder. There, aluminium loses all electrical resistance and fragile quantum states survive long enough to compute.
Each plate is colder than the last: about 50 K, 4 K, 1 K, 100 mK and 10 mK. Every stage intercepts heat and noise so the next can go further down.
Control pulses travel down from room temperature through dozens to hundreds of coaxial lines, attenuated stage by stage so thermal noise never reaches the chip. Faint answers come back up through amplifiers parked at 4 K.
At the very bottom hangs a gold-plated box. Shielded from light, heat and stray magnetic fields, it holds the only part of this whole machine that actually computes.
The copper lid lifts away. Underneath is a silicon chip about a centimetre across, wire-bonded to a circuit board. The bright metal is superconducting aluminium, patterned like any microchip. Except every trace is a quantum circuit.
These are transmons: tiny superconducting circuits that behave like artificial atoms. Their energy levels are discrete, so the lowest two can stand in for 0 and 1. Thin couplers let neighbouring qubits interact.
A few tens of nanoseconds of microwave light rotates a qubit. Resonators along the chip edge read its state out. String enough pulses together in the right order and you have a quantum algorithm.
The glowing loop is a Josephson junction: two aluminium layers separated by an oxide barrier only a nanometre or two thick. It makes the energy levels uneven, so microwaves can address just the lowest two. That is your qubit.
A bit is 0 or 1. A qubit can be any point on the Bloch sphere: north pole is 0, south pole is 1, everything between is a superposition. Measure it and it collapses to a pole, with odds set by how close it was.
Entangle two qubits and each one alone looks perfectly random, its arrow shrinking to nothing. The information lives between them, in correlations no local classical model can reproduce.
With many qubits, an algorithm choreographs amplitudes so wrong answers cancel and right answers reinforce, like ripples on a pond. That interference is where quantum speed-ups come from, for the right kinds of problems.
QUBIT LAB · LIVE SIMULATION
This is a real single-qubit simulator. Each gate rotates the arrow on the Bloch sphere, exactly like a microwave pulse rotates the qubit inside the machine you just took apart. Drag the sphere to look around.
Pick a gate. Every gate is a rotation.
Measure the qubit to build a histogram. Every measurement is genuinely random.
circuit: (empty)
WHERE IT MATTERS
Today's machines have hundreds of noisy qubits. Useful, error-corrected ones are still being built. That is exactly why the teams who understand the technology now will be the ones ready to use it.
Molecules are quantum systems, so simulating them is a natural fit. Better catalysts, batteries and drugs are the long-term prize.
Routing, scheduling and portfolio problems explode in size. Quantum and hybrid heuristics are being tested against the best classical solvers.
A large fault-tolerant machine could break today's public-key encryption. Migrating to post-quantum cryptography is already under way.
Quantum kernels and generative models are an active research area. Whether they beat classical methods on real data is still an open question.
Quantum amplitude estimation promises a quadratic speed-up for Monte Carlo methods used in pricing and risk analysis.
Atomic clocks, magnetometers and gravimeters already use quantum effects in the field to measure time, fields and motion with extreme precision.