Climate startup pilots: a three-clause contract fix

A multinational chemicals firm wants to trial your thermal battery system at one of their plants — three months, on-site installation, a custom integration with their legacy control system. The procurement email is friendly. The pilot agreement is twenty pages and reads like a partnership proposal: no fee, no quantified success metrics, and a clause granting the client joint ownership of any data the system generates. Sign in two weeks, or they go to a competitor.

This pattern shows up again and again across industrial heat, agritech, building retrofits, and EV charging deployments. Enterprise pilots in heavy climate sectors routinely fail not because the technology underperforms but because the contract treats a deployment as a favour rather than a commercial transaction. The hardware gets installed. The data flows. The startup writes off a quarter of engineering capacity. Then the procurement contact rotates, the plant manager forgets your name, and the pilot becomes "ongoing evaluation."

That is pilot purgatory, and it kills climate hardware businesses faster than bad engineering ever will.

Escaping the Pilot Purgatory: Why Informal Trials Stall

The phrase "let's run a pilot" sounds collaborative. In practice, in the climate industrial stack, it is usually the start of a negotiation where the enterprise has all the leverage and the founder has all the enthusiasm. The dirty secret of climatetech enterprise sales is that pilots are not free market research for the buyer; they are heavily discounted custom engineering for the seller, dressed in the language of partnership.

Three structural reasons pilots stall:

• Unstated or shifting success criteria. "Let's see how it performs" is not a metric. When the trial ends, the buyer can redefine success to match whatever outcome emerged, and the seller has no contractually anchored defence.
• No budget signal. A pilot without a fee signals one thing inside the enterprise: this is not a real project. No internal stakeholder has had to defend a budget line. When the procurement contact changes or the plant manager's priorities shift, there is no commercial gravity holding the trial in place.
• Ambiguous ownership of the learnings. The hardware generates terabytes of operational data — the most valuable asset a climate startup produces. If the contract grants the buyer rights to that data, or leaves ownership unclear, the founder may have effectively subsidised the customer's R&D roadmap.

These three failure modes are not accidental. They are the default settings of informal trial culture, and they persist because they benefit the enterprise side. The fix is not better relationships. The fix is contract architecture.

A pilot agreement is not a handshake that happens to be written down. It is a commercial instrument that signals — inside both organisations — whether this is real.

Defining Success: Quantifiable Metrics vs. Vague Expectations

The first clause that separates a real pilot from a trial that drifts is the success metrics clause. And here the trade-off is sharp: enterprise buyers want flexibility, founders want certainty, and the contract has to thread both.

Three things have to be present in a quantified success clause.

A locked baseline measurement. Before the hardware is switched on, both parties agree on the current state of the system being replaced — kWh of thermal energy consumed, tonnes of CO₂ emitted per production batch, hours of downtime per quarter. The baseline goes in writing, ideally with a methodology annex attached. The Chancery Lane Project, which publishes open-source climate-aligned contract clauses, treats this kind of measurement anchoring as table stakes for any serious decarbonisation contract; their supply contract clauses recommend a three-month deadline for setting emissions reduction targets and a six-month window for the corresponding transition plan to be operational.

A measurable outcome. Not "we'll evaluate performance." A number, with a tolerance band. "Reduce natural gas consumption in the curing process by 18–22% over a 90-day rolling average, measured at the meter specified in Annex B." If you cannot get a number, you do not have a pilot — you have a demo.

An arbitration path for disputed outcomes. What happens if the buyer says "your system performed at the lower end of the band, so we're calling it inconclusive"? The clause should specify how that disagreement gets resolved — independent measurement, a pre-agreed third-party reviewer, a defined period for counter-evidence. Without this, the success metric is a suggestion.

The emotional trap for founders is that pushing for numbers feels adversarial. It is not. Quantified success is what allows the buyer's internal champion to defend the project upstairs. Without a number, that champion has nothing to bring to the board. The contract is doing them a favour by forcing clarity.

The Financial Signal: Why Unpaid Pilots Lack Commercial Commitment

The second clause is the one founders avoid most, because it feels like the dealbreaker: payment. But an unpaid pilot is not a discount. It is a missing signal.

Inside an enterprise, budget is how seriousness is communicated. When a buyer says "we'll pay nothing for the pilot," what they are telling their own organisation is that this project does not need defending. No line item. No procurement review. No CFO sign-off. The plant manager who championed the trial can walk away without consequence, because no one lost money when it ended.

For climate startups, the cost is asymmetric. Custom integration work, on-site engineering time, hardware retrofitting — these eat runway in a way that customer discovery calls never do. When the pilot ends without conversion, the founder has effectively funded a competitor's evaluation, and the team's attention is locked for a quarter, unable to pivot toward a buyer with real budget gravity.

So what does a paid pilot look like in practice? A few patterns work:

• A discounted pilot fee that reflects the reduced risk to the buyer but still requires a purchase order. Even a modest fee — covering integration costs plus margin — forces the buyer to put the project through real internal processes.
• A milestone-based structure. The buyer pays for each phase as it completes: site survey, installation, baseline measurement, performance period. This keeps cash flow manageable for both sides and creates a natural cadence for go/no-go decisions.
• A conversion right priced into the pilot fee. The pilot fee is credited against the first year of a full commercial licence or supply agreement, so the buyer's pilot spend is not lost if they convert. This is closer to how venture-style investors structure things; EIT Climate-KIC's Climate-SAFE programme, for instance, invests €50,000 in cash and services in climate startups with a 24-month long-stop date and converts when the company raises a qualifying funding round above €250,000. The structure is identical in spirit: a real commitment now, a defined trigger later.

The hard trade-off is that some pilots will die at this clause. That is the point. A pilot that cannot survive a paid structure was never going to convert anyway. Better to know in week three than month nine, and to redirect engineering attention toward the buyers who actually have budget gravity.

Protecting Your IP: Data Ownership and Indemnification Essentials

The third clause is the one founders learn to care about only after they have lost something. Data ownership and indemnification sit in the back of the contract because they look like legal hygiene. They are not. For a climatetech hardware founder, they are the core asset.

Two questions have to be answered unambiguously in the pilot agreement.

Who owns the operational data the system generates? The default in many enterprise contracts is joint ownership, or — worse — sole ownership by the buyer. For a thermal battery or a precision fermentation platform or an EV charging optimisation system, that data is the moat. It is what makes the next deployment faster, the next model better, the next customer more confident. If the buyer owns it, the founder has effectively trained the customer's in-house team on the founder's own technology.

The contract should specify that the startup retains ownership of aggregated, anonymised operational data, and grants the buyer a licence to use its own site-specific data for internal reporting. Some buyers will push back. The right answer is to offer a defined data-sharing window — twelve months of anonymised insights for the buyer's ESG reporting, say — in exchange for retention of the underlying IP.

Who carries the risk if a third party claims the technology infringes their IP? This is the indemnification clause, and it is where standard SaaS templates fail climatetech founders. Software vendors can carry broad indemnification because their surface area is small and well-mapped. A hardware deployment on a customer site — connected to their control systems, processing their materials, integrated with their legacy infrastructure — has a much wider exposure profile. If the buyer demands that the startup indemnify them against any IP claim arising from the deployment, including claims related to the buyer's own modifications or integrations, the founder is taking on risk they cannot price.

The clause should be reciprocal and scoped. Each party indemnifies the other for IP infringement claims arising from its own technology. The startup does not indemnify the buyer for the buyer's own modifications. The buyer does not indemnify the startup for misuse of the system post-handover. This is standard in technology licensing but routinely omitted from pilot agreements because both sides assume the trial is too small to litigate. By the time it is not small, the contract has already been signed.

From Pilot to Contract: Implementing Binding Conversion Terms

The final piece of architecture is what happens at the end of the pilot. Most pilot agreements fudge this — "the parties may negotiate a commercial agreement upon successful completion of the pilot." That sentence is pilot purgatory in legal form.

A binding conversion clause has three components.

A defined decision window. Within a fixed number of days of pilot completion, the buyer must either execute a commercial agreement on pre-agreed headline terms, extend the pilot for a defined period at an agreed fee, or pay a wind-down fee covering the startup's cost of decommissioning. No silent "we'll think about it" option.

Pre-agreed commercial terms. Not the full contract — but the headline economics. Price per unit, minimum volume, term length, payment terms. Climate-aligned contract drafting, the kind The Chancery Lane Project publishes, works toward exactly this kind of pre-agreed framework, embedding emissions reduction and reporting obligations into the commercial structure from day one so the pilot is a step toward a known destination rather than an open-ended experiment.

A documented handover if the buyer walks away. The startup retrieves its hardware, the buyer returns or destroys site-specific data, and both parties sign a short close-out memo that prevents the buyer from later claiming the pilot results for an internal project. This is unglamorous and often skipped. It is also the difference between a clean break and a six-month email chain about who owns the dataset.

The trade-off here is the founder's fear of looking transactional. Climate founders, especially in deeptech, often feel they are selling a mission as much as a product. The pilot conversation gets framed as collaboration. The contract reflects that framing — open-ended, friendly, light on obligations. But the enterprise side does not extend the same generosity when its own procurement team reviews the document. The asymmetry is structural, not personal. The way to close it is to write the conversion terms the way a serious counterparty would write them, even if it feels cold.

The pilot agreement is the first contract. If you would not accept its terms at scale, do not accept them for thirty days of evaluation.

The Hard-Earned Lesson

There is a version of the climatetech founder journey where pilots are a necessary evil on the road to revenue. There is another version where pilots are the business — a treadmill of free deployments that look like progress and deliver none. The difference is not the technology, the market, or the timing. It is the contract.

Three clauses. Quantified success with a dispute path. A paid commitment that signals internal seriousness. Clear IP and indemnification boundaries that protect the asset the founder actually owns. A binding conversion path that turns the pilot into a step rather than a destination. None of these are exotic. All of them are routinely omitted.

The messiness of building a climate company is not going away. Pilots will remain the primary commercial on-ramp for hardware-heavy climatetech for years. The resilience that matters in this sector is not the resilience to keep grinding through unpaid trials — it is the resilience to walk away from a bad deal on Tuesday so the right deal has room to land on Thursday. The founders who survive the pilot stage will be the ones who treat the pilot agreement as the most important document they sign all year, because it is. The ones who keep sending out friendly handshake agreements with vague KPIs and no fee will keep writing off engineering quarters to enterprises that never had to defend the spend.

Pick the clause. Pick the fight. The alternative is the purgatory, and it has no exit ramp.