Skip to main content

Loading interactive tools...

Quantum Platform · Early Access

The ultra-scale engine for yourquantum stack

Simulate up to 47 qubits with 3.6x memory compression across any GPU cluster. Deploy anywhere - from single-node cloud to multi-thousand GPU clusters.

Get Early Access
AmpliGrid distributed statevector capacity ceiling showing scaling from 4 to 47 qubits across 1 to 825 nodes, norm fidelity 0.9999999658

[ HARDWARE VALIDATION ]

Validated Performance

Empirically verified on production clusters. Zero extrapolation, zero theoretical estimates.

47QUBIT STATEVECTOR CAPACITY
825ISOLATED GPU NODES (3,300 GPUs)Validated on Isambard-AI
N is not equal to 2 to the power of mFLEXIBLE NODE ALLOCATION
99.87%Decoder accuracy
42+Chemistry modules
271Built-in algorithms
20+Quantum ML models

[ DISTRIBUTED ENGINE ]

AmpliGrid: distributed quantum compute, without the usual constraints.

Run quantum simulations across any GPU allocation without limits. No power-of-2 restrictions, supported in f32, f64, and memory-optimized block float precision.

AmpliGrid distributed benchmark performance chart

[ Platform capabilities ]

A complete stack. Simulation to discovery.

Four integrated core capabilities built directly on top of the AmpliGrid engine. Native execution across quantum ML, chemistry, drug discovery, and error correction. Zero wrappers.

Neural decoder

Pre-trained convolutional decoder for QLDPC surface codes, with real-time syndrome decoding and live confidence readout. Six error mitigation strategies included.

99.87% accuracyReal-time

Autonomous drug discovery

Describe a molecular target in plain language. The campaign engine automatically orchestrates docking, free energy calculations, and quantum simulations from goal to candidate.

Natural language goalsFull pipeline

20+ quantum ML models

QNN, QGAN, QCNN, QTransformer, quantum reinforcement learning, and more. All implemented natively without external ML frameworks.

20+ models10 optimisers

42+ chemistry modules

CASSCF, DMET, ADAPT-VQE, free energy perturbation, protein docking, and ADMET prediction: a complete pipeline from molecule to candidate.

42+ modulesADMET · FEP

[ Demos ]

Interactive Primitives

Drag, click, and run your way through quantum states, error correction, and circuit execution.

Bloch sphere

Drag to rotate · Click the sphere to place a state · Try the presets

|0⟩ North Ground state (|0⟩)
|1⟩ South Excited state (|1⟩)
|+⟩ East +X superposition
|i⟩ West +Y superposition (or +i)
ψ Tilted Arbitrary pure state

Neural decoder

Click dots to toggle stabiliser violations · Watch the confidence update

No errors All-zero syndrome -- no corrections needed
Single bit flip One stabiliser violation at position 12
Multi-bit error Three simultaneous syndrome violations

Circuit visualiser

Click a gate to inspect · Press Run to execute

4 Bell state Apply H + CX to create a Bell pair
6 GHZ state Extend entanglement across three qubits
8 Teleportation Transfer state using entanglement and classical comms

[ Algorithm library ]

271 built-in circuit templates

From state preparation to tensor networks: everything research-ready.

Early access · limited cohort

Ready to go further?

Request early access and start running real quantum workloads on AmpliGrid, validated at 47 qubits on Isambard-AI GH200.

Request Early Access
47 Qubits validated Isambard-AI · NVIDIA GH200