Every sovereign compute project lives or dies on the same question: who pays the electricity bill on the Tuesday after the ribbon is cut. A quantum facility is no different. The hardware is novel, the operators are scarce, and the cooling stack is unforgiving — but the commercial logic is ordinary. You need workloads with real budgets and real urgency, run by buyers who can sign. For Ireland Quantum, committed for delivery in Q2 2027, three workloads do that job: pharmaceutical discovery, materials chemistry, and a narrow band of sovereign defence research. The rest of the book — finance, logistics, climate modelling — is welcome, and will come, but it does not anchor the build.
Why these three, and not the usual list
Most quantum prospectuses lead with portfolio optimisation in finance. It reads well. It also competes directly with classical solvers that are very, very good and getting better every quarter. The honest near-term advantage of a gate-based or analogue quantum machine sits in problems where the underlying physics is quantum — molecular energies, electron correlation, lattice behaviour, spin systems. That is chemistry. That is materials. And, in a narrower way, that is the simulation work that sits behind sensing, propulsion, and cryptographic research inside national defence programmes.
This is not a marketing point. It is what the machines are actually good at first. If you align the commercial pipeline with the physical advantage, you get paying customers in year one. If you don't, you spend three years explaining why the benchmark didn't beat a laptop.
Workload one: quantum drug discovery for Irish and visiting pharma
Ireland already hosts a large share of the world's pharmaceutical manufacturing. The discovery work is more distributed — some sits here, much sits in Boston, Basel, Cambridge — but the relationships are dense, the regulatory comfort with Irish suppliers is high, and the talent pool in computational chemistry around the Irish universities is real.
The workload itself is narrower than the slogan. Quantum drug discovery, in 2027, will not design a new oncology asset end-to-end. What it will do is supply tighter ground-state energy estimates for small but stubborn molecular fragments where classical density functional theory is known to be wrong or expensive. Think transition-metal active sites in catalysts, or the conformational energy of a ligand near a binding pocket where the standard methods disagree by more than the decision threshold.
A typical engagement looks like this. A computational chemistry team inside a pharma sends a list of fragments and the questions they want answered — usually a small set of energies, a barrier height, sometimes a spin gap. We agree a problem encoding, run the variational or phase-estimation routine on the machine with classical pre- and post-processing, and return numbers with error bars. The deliverable is a report a chemist can paste into an internal review. The contract is for compute time plus applied-science hours, billed against a defined work package. It is, in commercial shape, very close to how the pharma industry already buys CRO and HPC services. That familiarity is the point. We are not asking buyers to invent a new procurement category.
Workload two: quantum materials science for the people who actually make things
The second workload is quantum materials science, and it is the one I am most personally interested in. Battery cathode chemistry. Catalysts for green hydrogen and ammonia. Superconducting and magnetic materials for the next generation of sensors. High-entropy alloys for aerospace. Photovoltaic absorbers beyond silicon. The list is long because the gap between what classical simulation can predict and what an experimentalist actually finds in the lab is still wide, and every percent of that gap closed is worth real money to whoever closes it first.
Irish industry has more skin in this than people credit. There are world-class materials groups in Cork and Dublin, a serious medtech cluster around Galway, and a quiet but substantial network of speciality chemical and ingredient companies across Tipperary, Limerick, and the south-east. None of them will operate a quantum machine in-house. All of them, given a sane interface, will buy time on one.
The engagement model here is slightly different from pharma. Materials customers tend to want longer, exploratory campaigns — six months of structured screening rather than a one-shot energy calculation. So we sell it as a campaign: a fixed scope, a quarterly review, a defined exit point at which the customer either commits to a follow-on phase or walks away with what they have. That structure protects both sides from the dominant risk in quantum, which is not that the machine fails, but that the question turns out to have a cheaper classical answer halfway through.
Workload three: sovereign defence, on Irish terms
The third workload is the one that requires the most care to talk about, and I will be plain. Ireland is militarily neutral. That is a constitutional and political fact, and Ireland Quantum is being built inside it, not around it. Sovereign defence quantum work, in this context, means a narrow set of permitted research activities for the Defence Forces, for European partners under existing civilian-research frameworks, and for adjacent agencies whose remit is communications security, sensing, and resilience rather than weapons.
The work is mostly simulation and algorithm development. Post-quantum cryptography validation against realistic attack circuits. Sensor modelling for navigation in environments where GPS is denied or degraded. Quantum-enhanced timing and synchronisation for critical national infrastructure. Materials work for radiation-hard electronics. None of this is exotic. All of it is the kind of thing a small, neutral, technologically serious country should be able to do for itself rather than rent from somebody else's stack.
The commercial significance is that defence-adjacent budgets are stable, multi-year, and contracted in a way that smooths the lumpiness of commercial pharma and materials work. That stability is what lets you offer pharma the long appointment times they want without the operations team panicking about utilisation in the off-quarters.
What "quantum as a service" actually looks like on a Tuesday
The phrase quantum as a service has been worn smooth by overuse, so let me describe what it concretely is at our facility. There is an API. There is a queue. There is a portal where authorised users from a contracted organisation can submit jobs against their allocation. There is a small applied-science team — physicists, computational chemists, software engineers — who sit between the customer and the metal, because in 2027 you cannot honestly hand a chemist a raw gate-level interface and expect a useful result on Tuesday.
So a typical engagement runs in four steps:
- Scoping. Two to four weeks. We sit with the customer's domain team, look at the problem, and decide whether quantum is the right tool. About a third of the time, the honest answer is "not yet, run it classical." We say that out loud. It is the single most important thing for the long-term reputation of the facility.
- Encoding. The applied-science team translates the chemistry or materials problem into a circuit, an ansatz, an embedding — whatever the relevant abstraction is for the workload and the hardware generation we are running.
- Run. Compute time on the machine, with classical pre- and post-processing. This is the part the customer is technically buying. It is rarely the part that takes the longest.
- Report. Numbers, error bars, a written interpretation, and a clear statement of what the result does and does not allow the customer to claim. The last sentence matters more than the first.
Pricing is per-engagement, not per-shot, because per-shot pricing punishes good experimental design. Customers who buy time in volume sign annual access agreements with a guaranteed allocation. Smaller customers — university spinouts, single-product biotechs — buy single work packages.
The honest constraints
I want to name the three things that will go wrong, so nobody is surprised when they do. First, the hardware in 2027 will not be the hardware in 2030, and some early customer engagements will produce results that are superseded inside eighteen months. We tell customers this in the scoping call. Second, the talent market for people who can run these systems is global and tight, and we will lose people to better-funded labs unless we keep the work interesting. Third, the gap between quantum chemistry as a research field and quantum chemistry as a billable service is real, and bridging it is operational work — documentation, tooling, repeatability — that does not look glamorous on a slide but pays the rent.
What to do this week
If you run a research function inside an Irish pharma, materials, or speciality chemicals business and any of the above sounds like a problem you currently work around rather than solve, write to us. We are taking expressions of interest now for the Q2 2027 cohort, and the scoping conversations we are having in 2026 are the ones that decide who gets allocation in the first year. The build is funded against the workloads. The workloads are decided by the customers who turn up early. That is the whole mechanism. Everything else is decoration.