What Is FOAK in Climate Tech? From Pilot to First Commercial Deployment

This guide defines FOAK and explains why the first commercial plant is a different decision. For why the first plant is hard to finance, read Why Are FOAK Climate Projects Hard to Finance? Risk, Ambiguity and the Missing Reference Class.

In April 2024 Florian Tiller, founder of the Berlin direct air capture startup Ucaneo, told us in an interview that customers across several industries were already asking him one question. “When you're ready, what costs will you have?” Ucaneo had “quite a big running prototype in the lab”, a third-party feasibility study behind it and a target cost below €300 per tonne of CO2 within three years. The next step was a standardised industrial unit. In the same conversation he named the question his whole category has to answer. “If you look out there, every direct air capture technology will always tell you they’re the cheapest, the best, and the most energy-efficient. But the question is always at which scale do you achieve that? Is it at a million-tonne scale? Or a 500-tonne scale? And by when?”

The €300 figure was a 2024 target, and the interview records one founder's account of where his company stood at the time. The question he put to his own industry is the FOAK question. A lab prototype can show that a mechanism works. Only a built commercial asset can tell a customer what the product costs at a stated scale and by a stated date.

What FOAK means in climate tech

FOAK stands for first-of-a-kind. In climate technology it usually describes the first commercial-scale deployment of a technology or configuration after laboratory, pilot and demonstration work. The boundary varies by sector. A project can be first at a particular scale, first in a geography, first in a configuration or the first commercial use of a technology that has already run elsewhere. The novelty can sit in the core process, the feedstock, the site integration, the product specification, the equipment configuration or the commercial arrangement, and each of those is a different kind of first.

That flexibility makes the label useful only when the company explains what it means. Two projects can both be called FOAK while carrying very different levels of novelty. One may use proven equipment in a new process integration. Another may combine a new process, new feedstock, new site and a buyer relationship that has never existed before.

Climeworks shows how one plant becomes the reference for the next

Direct air capture also gives a public answer to Tiller's question one stage later. Climeworks launched Orca, its earlier and smaller plant in Iceland, and then broke ground on Mammoth in June 2022. Mammoth began operating in May 2024 and is designed for a nameplate capture capacity of up to 36,000 tonnes of CO2 a year. Climeworks is explicit that actual net removal will be lower than nameplate capture once the full production waterfall, its term for the losses between gross capture and net removal, is applied. Public reporting in 2025 put Mammoth’s captured volume well below nameplate, and in September 2026 the company reported a sixfold increase in capture.

That detail is what a FOAK description should preserve. Nameplate capacity, net removal and delivered tonnes are different facts, and each answers a different part of the question about scale. Mammoth matters because it gives Climeworks engineering and operating experience for later scale-up, and because every later direct air capture plant will be read against its record.

The decision changes when the asset becomes commercial

A pilot establishes whether the technology can work under relevant conditions. A commercial project has to show that a particular asset can be built at a particular site, integrated with the surrounding system, operated to specification and paid for by real counterparties. Buyers, lenders, engineering contractors, host sites, insurers and public funders each read the project through their own mandate. Pilot data remains important, and it now sits inside a larger decision about contracts, delivery, risk allocation and capital.

The IEA has described clean energy innovation as having a commercialisation problem, and the US Department of Energy's Pathways to Commercial Liftoff work deliberately combines technical readiness with market formation, investment, demand and deployment. FOAK is where those conditions meet one balance sheet.

The capital changes size at the same moment, which our guide to why FOAK climate projects are hard to finance covers in depth. The company moves from financing a technology company to financing a company and a specific asset at once. That puts the boundary of the asset inside the definition. A reviewer needs to know what is being built, which infrastructure is included, what belongs to the host site and which costs sit outside the project company. The test fits on one page. Draw the boundary, place each cost inside or outside the project company and name the host's share. This matters most when a venture deck is reused to raise project capital, because the company can be compelling while the asset is still undefined.

FOAK sits between demonstration and repeat deployment

The words used around scale-up vary by sector and funder. The sequence itself is broadly recognisable. FOAK is the first commercial-scale project with real contracts and economics. J.P. Morgan's 2026 climate-tech work uses the term in the same position, between demonstration and repeated commercial deployment. Projects that follow are often described as early-of-a-kind while the design still changes quickly, and Nth-of-a-kind, or NOAK, refers to later projects that can draw on a genuine record of built assets.

These are working categories. A direct air capture facility, fermentation plant and thermal-storage installation do not share a single definition of commercial scale, which is why the useful description is the evidence itself. State what has been demonstrated, at what scale, for how long and under which operating conditions, then name what the next deployment still has to prove. Evidence has an envelope. A pilot that ran for a thousand hours on one feedstock tells a reviewer about uptime at that scale, on that feedstock and in that operating range. Ten times the throughput on a different feedstock and duty cycle sits outside it. Write down the largest relevant system already operated, for how long and under which conditions, and mark which elements of the new project fall outside that envelope.

A high Technology Readiness Level (TRL) says nothing on its own about offtake, permitting, unit economics or bankability. A simple test is that every TRL quoted in the materials carries the status of supply, offtake, permits and financing beside it. A blank field counts as a fail.

1. 01Lab and prototypeThe mechanism can work.
2. 02PilotIt operates outside the lab at limited scale.
3. 03DemonstrationAn integrated system runs close to commercial conditions.
4. 04FOAKThe first commercial-scale project, with real contracts and economics.
5. 05NOAKLater plants draw on a reference class of built projects.

Labels vary by sector and funder. What each stage proves stays the same.

The useful question is which parts are actually first

A project can be novel in one dimension and familiar in several others. The core process may be new while civil works are conventional. The product may meet specification while no buyer has purchased it at the planned volume. So name the novelty precisely. “First commercial deployment of this process at this throughput with this feedstock” gives a reviewer something to test. “FOAK technology” leaves the whole project inside one undifferentiated risk label.

Some investors look for designs that keep the novel part small. Carlota Ochoa Neven Du Mont of the climate fund Extantia Capital said in her October 2024 interview that “If you can scale by connecting multiple smaller units instead of building larger plants, it reduces complexity and speeds up deployment.” That is one fund's preference, stated for hardware it backed in 2024, and it shows how an investor reads novelty as something the design can shrink.

The parts that stop a first project are often non-technical. Permitting, grid access, land rights, construction capacity, feedstock, product qualification, insurance and the financing timetable can each become the controlling constraint while the core process works. A credible description shows which of these conditions are closed, which have a defined pathway and which remain open. Breaking the project into parts and marking where precedent exists turns a single FOAK label into a list a financier can assess. Our piece on why nobody wants to become the precedent covers why investors often prefer to back the company that comes second.

Brighter Future will not turn the word FOAK into a badge of innovation on a deck. The label earns its place only when the reader can see what is first, what is proven and what the next project should learn from this one.

When a company prepares the case for a first plant, its finance team owns the figures, and our investor storytelling work builds one evidence register that marks each one contracted, quoted, modelled or assumed.

Project one has to leave a record for project two

A first plant makes product. It also produces the only data anyone will have on the technology at commercial scale, and the projects that follow are priced on that data. Lenders price first projects expensively because they lack a reference class, a set of comparable built projects whose cost and performance they can check. Project one is the chance to build it. If nobody decides in advance what to record, the learning scatters across contractors, operators and inboxes. The next investor then receives anecdotes, project two is priced as another first-of-a-kind, and the company pays the premium for being first twice.

The record has four parts. Technical performance covers throughput, yield, uptime, efficiency, degradation, product quality and energy use across the full range of operating conditions. Delivered cost and schedule show where cost moved against the pre-construction model and which tasks, long-lead items, permits or interfaces controlled the timeline. Customer acceptance shows how often product fell off specification, whether buyers took their contracted volumes, which price assumptions held and what objections appeared once real product arrived. Operations covers staffing, skills, maintenance, failed parts, near misses and the tasks that still depend on founder knowledge nobody has written down.

The most useful output is a split between costs that belonged to being first and costs that will recur. First-time engineering, redesign during construction and the time a team spends learning a new build go in the first column. Equipment bought for every plant, standard civil works and routine operating labour go in the second. That split lets an investor believe a cost curve, and it shows which parts of the design can become standard modules and which stay site-specific.

Define the record before construction starts. For each claim the company expects to make about project one, log the expectation before construction, the observed result, the source and date, the quality and limits of the evidence, why it matters to project two and the approved external wording. Kept from the first day, that log lets the project-two pitch say what was uncertain before the first plant, what is now known, what changed in cost, schedule, performance and demand, and what remains site-specific. If project two goes out with the same claims and the same uncertainties as project one, the first asset created too little evidence for the company to carry forward.

The first plant answers the customer's question

Tiller's customers were asking a question that no prototype could settle. A cost per tonne becomes a fact only when a plant of stated size has run long enough to produce one, and a FOAK project exists to make that fact. Its delivered cost then does a second job. It becomes the number the second plant's buyers, lenders and engineers judge that plant against, so the first answer to “what costs will you have?” sets the terms of every conversation that follows.

Sources and further reading

• MaRS launches Canada's first-of-a-kind lab to scale cleantech champions, MaRS Discovery District, 2026
• Energy innovation has a commercialisation problem, IEA
• Pathways to Commercial Liftoff, US Department of Energy
• Climate-tech industry trends, J.P. Morgan, 2026
• Mammoth, Climeworks
• The carbon removal production waterfall, Climeworks
• Technology readiness level definitions, NASA
• Florian Tiller, Ucaneo, founder interview, April 2024
• Carlota Ochoa Neven Du Mont, Extantia Capital, investor interview, October 2024
• Why Are FOAK Climate Projects Hard to Finance?
• Why Nobody Wants to Become the Precedent