The full timeline — Q3 2026 through Q4 2027

What we are actually building, and when

The Ireland Quantum 100 timeline runs from Q3 2026 through Q4 2027. Five quarters, four engineering gates, one operational machine: a 100-physical-qubit superconducting transmon system housed in Co. Tipperary, with climate workloads as the priority cohort. This page is the full schedule with the engineering context behind each gate — what gets installed, what gets characterised, what we expect to learn, and what we do not expect to have working at each point.

The honest framing first: a 100-qubit transmon device in 2027 is not a fault-tolerant machine. It is a noisy intermediate-scale (NISQ) platform with a credible path toward distance-3 surface-code patches on a heavy-hex layout. Most of the milestones below are about getting the cryogenic, control-electronics, and software stack to a state where chemistry and optimisation workloads run end-to-end with defensible error budgets — not about declaring quantum advantage.

Q3 2026 — Site fit-out and cryogenic plant

The Tipperary fit-out is the unglamorous quarter that determines everything downstream. A dilution refrigerator targeting sub-15 mK base temperature is unforgiving about its building. The room needs vibration isolation below roughly 1 Hz (HVAC, road traffic, footfall all matter), magnetic shielding sufficient to keep ambient field drift well under a milligauss at the mixing chamber stage, and a power feed clean enough that compressor pulse-tubes don't beat against the readout chain.

• Reinforced concrete plinth on isolated foundations, decoupled from the surrounding slab
• Mu-metal and aluminium magnetic shielding around the fridge enclosure
• Dedicated helium plant — pulse-tube precooling plus mixed 3He/4He circulation for the dilution stage
• Filtered, regulated three-phase supply with UPS-backed control racks
• Class-7 cleanroom adjacent to the fridge for sample handling and wirebonding
• RF-shielded screened room for the control electronics, with phase-stable cabling runs to the fridge head

By end of Q3 2026 the building is the instrument. No qubits yet — just an empty cryostat capable of holding millikelvin temperatures stably for weeks at a time. We will publish the thermal stability and vibration spectra as part of the commissioning record.

Q4 2026 — Cryostat install and control-stack integration

The fridge gets populated. This is the quarter where the heavy-hex chip carrier, attenuator chain, circulators, isolators, travelling-wave parametric amplifiers (TWPAs), and HEMT amplifiers all go in, stage by thermal stage. A typical control chain looks like roughly 60 dB of attenuation distributed across the 4 K, still, cold-plate, and mixing-chamber stages on the input side, with TWPA + HEMT + room-temperature amplification on the readout side.

The control electronics are arbitrary waveform generators driving microwave drives in the 4–8 GHz band for transmon control, with separate flux-bias lines for tunable couplers. Software stack is OpenQASM 3 at the program layer, with Qiskit, PennyLane, and Cirq supported as front-ends compiling down to a native gate set on the heavy-hex topology. The pulse-level layer is where the real work happens: DRAG-shaped single-qubit gates, calibrated cross-resonance or tunable-coupler two-qubit gates, and dispersive readout pulses optimised against the resonator linewidth.

Gate criterion at end of Q4 2026: cold cryostat with a populated chip, full control-line connectivity verified at room temperature, and a documented calibration plan for first-light.

Q1 2027 — First light: single-qubit characterisation

"First light" in the quantum facility milestones sense means the first qubit responds coherently to a drive pulse and we can extract T1, T2, and single-qubit gate fidelity numbers we trust. Expected sequence:

• Resonator spectroscopy — locating each readout resonator and its dispersive shift
• Qubit spectroscopy — two-tone measurements to find the 0→1 and 1→2 transitions and extract anharmonicity
• Rabi, Ramsey, and Hahn-echo sequences to pin down T1 and T2*
• Randomised benchmarking on single-qubit Cliffords for per-gate error
• Readout-error characterisation and discriminator training

We are targeting a working subset of the 100-qubit array by end of Q1 2027, not the full array — yields on superconducting devices are real and we will not publish a fabricated number ahead of measurement. The deliverable is a calibrated, characterised set of qubits with published coherence and gate-fidelity data, plus the calibration cadence required to keep them stable.

Q2 2027 — Multi-qubit operation and external access

This is the milestone the Ireland Quantum 2027 conversation has been pointing at: two-qubit gates across the heavy-hex lattice, and the first external workloads. Engineering deliverables:

• Calibrated two-qubit entangling gates on neighbouring pairs across the heavy-hex graph
• Cross-talk characterisation and simultaneous-RB across overlapping pairs
• A native compiler pass that maps logical circuits to the heavy-hex topology with SWAP-aware routing
• Error-mitigation primitives: zero-noise extrapolation, probabilistic error cancellation, dynamical decoupling on idle qubits
• A small distance-3 surface-code patch as the first concrete step on the error-correction roadmap — purely as a characterisation experiment, not a logical-qubit claim
• External-user access via a queued job API, OpenQASM 3 in, results out, with transparent calibration metadata attached to every job

The first external cohort is climate-science: variational quantum eigensolver (VQE) runs on small carbon-capture amine and metal-organic framework fragments, photovoltaic absorber candidates, lithium-sulphur and solid-electrolyte battery chemistries, and grid-optimisation problems formulated as QUBO instances for QAOA. These workloads feed directly into IMPT.io's offset-stack supplier candidate evaluation — the chemistry running on the machine is the chemistry that determines which carbon-capture and materials suppliers we can credibly recommend.