Not Engaging EPCs for Climate Projects, Part 2: When the Traditional Model Isn’t the Right Fit

Lessons from Enduring Planet & Friends' Financing Series

A full-service EPC can take a lot of execution risk off your plate. But for first-of-a-kind climate projects, that risk transfer can also mean higher costs, over-engineering, and a partner whose incentives do not always line up with yours.

At our latest Enduring Planet & Friends Financing Series session, “Engaging EPCs for Climate Projects, Part 2,” Hannah Friedman (Lupine Finance) and Dimitry Gershenson (Enduring Planet) led a conversation with Christian Okoye (Occam Edge), Will Phelan (Molten Industries) and Amit Modi (FGS) on when the traditional Engineering, Procurement, and Construction (EPC) model may not be the right fit, what founders take on when they choose another path, and how to manage the risks that come with it.

The takeaway: Going without a traditional EPC does not proportionally speed up projects or save costs compared to the amount of residual risk it leaves. It moves a disproportionate amount of risk onto your team. That can give you more control over design, procurement, and cost, but only if your team has the experience, balance sheet, and internal capacity to manage what the EPC would otherwise own.

Easy heuristics like “if you need to raise project finance, go find an EPC” still hold, but in the grey area the real question is: Which risks are you equipped to hold yourself, and which ones are better transferred to an experienced partner?

Start With Assessing a Project’s Profile

Not all first-of-a-kind risk is the same. A project built from familiar equipment in a new configuration may primarily carry some integration and operational risk. It’s possible a seasoned general contractor can manage this. A project relying on technology that has not operated commercially at the required scale carries meaningful technology and performance risk. That distinction should shape which contractor(s) you approach and, importantly, what they can realistically guarantee.

FOAK projects can create a natural incentive to over-engineer. An EPC being asked to stand behind a system with limited to no operating history at scale has good reason to design conservatively. More redundancy and engineering can reduce the contractor’s exposure to failure. 

While this may not align with the incentives of your developer, lender (or even your VC board) who may be trying to achieve the most attractive risk-adjusted return possible, sometimes over-engineering with the experience of an EPC creates higher certainty of delivery and of safety.

The trade-off is bankability. Project finance lenders want detailed engineering, credible cost estimates, schedules, contracts, performance protection, and clear risk allocation. And for good reason. That level of documentation may feel heavy compared with how an early-stage company normally operates, but it has been codified as the gold standard for building infrastructure repeatedly over many years.

In part, lenders seek this degree of confidence because if the EPC is not wrapping the risk, someone else has to. Project financiers generally want performance guarantees somewhere in the structure. If the contractor cannot provide them, or you choose to go without a full EPC entirely, the company may need to stand with its own balance sheet behind the project. Contractor selection is therefore a financing decision as much as a construction one.

There Are a Few Ways to Unbundle the EPC

Tier 1 might not suit, but an EPC could still be the right partner. Large EPCs bring deep experience and stronger balance sheets, but their economics are often geared toward much larger projects than an initial commercial demonstration. A Tier 2 EPC with relevant experience and greater willingness to collaborate may be a better operational fit for a FOAK project.

Incentives help, but they cannot fix a poorly structured project. Bonuses, shared savings, or schedule incentives can improve alignment, but they are not substitutes for clear specifications, design control, and experienced project management. Push too hard on speed and you can simply replace schedule risk with quality risk.

There is a middle ground between a full EPC and doing everything yourself. Progressive design-build can bring the EPC into the FEED process early so the developer and contractor shape the project together, improving construction input and pricing certainty before the design is locked. Another option is separating design from build, keeping more control over engineering while relying on experienced partners for construction.

Whichever model you choose, define ownership at every interface. The cost advantage of unbundling disappears quickly if no one is clearly responsible when scopes slip.

AI isn’t ready yet. It’s very tempting to consider how AI can disrupt and make the role of the EPC more efficient. Certainly, AI can support parts of the process, including organizing information and assisting with early analysis, but it can also create false confidence when used for critical financial modeling or engineering without deep subject-matter expertise. It does not replace experienced judgment around the risks that actually determine whether a project gets built. Remember: Big project construction and operation carries life-and-death decisions for the safety of builders and operators, and thoughtful EPCs have stated why these tools aren’t ready to shoulder that burden yet. 

The Long, Lonely Road Going It Alone: What to Watch

More control means more execution responsibility. Managing engineering, procurement, and construction separately can give you tighter control over specifications, equipment selection, and cost. It also means tackling the coordination work a full EPC would otherwise own.

Many of the hardest project problems happen at the interfaces: when engineering meets construction, when factory-built equipment has to integrate in the field, or when one contractor’s scope or timeline depends on another’s. Modularization can reduce some execution complexity, but it does not eliminate the need to manage those handoffs precisely.

If you go without a traditional EPC model, the mere process and management of that process can become hefty. Don’t underestimate the effort required to translate decisions from engineering to construction; this is a communication handoff that an EPC is well-steeped in.

Relevant, prior experience at a higher complexity than what you’re building now, matters.  If you intend to retain more execution risk, bring in people who have already delivered projects larger and/or more complex than the one you are building. This is perhaps one of the single-most important “readiness” metrics if you’re considering going your own way. 

Project development timelines move, and hiring aggressively against an optimistic schedule can leave you carrying a large team while engineering, permitting, or financing is still catching up. At the very outset, even before building a large team, build a detailed risk register alongside the financial model, then have experienced operators map how they will actively retire those risks. 

A generic contingency percentage can create false confidence. For FOAK projects, simply putting 10% or 20% into the model may not reflect where the actual risk sits. Contingency should be built package by package based on design maturity, execution complexity, labor exposure, and other uncertainties. Some less-defined packages can require substantially more cushion than the overall project average. Your contingency should come down because uncertainty has been resolved, not because the project excel model needs a better return.

Your schedule needs contingency too. Commercial-scale project development almost always introduces risks that were not visible at the technology demonstration stage. Integration, construction, permitting, procurement, and financing can all move the critical path. That means the corporate runway supporting the project needs to survive a development timeline that is longer than the base case, particularly when the company is carrying more of the execution scope itself.

Tactical Takeaways

There is no universally right EPC structure for a climate project. A full EPC can be expensive and conservative, but it transfers risk and can improve bankability. Unbundling the model can give founders more control and potentially lower costs, but it also puts more integration, performance, and execution risk back onto the company.

If you are deciding how much of the EPC process to own yourself, consider how you could:

• Match the contractor to the project’s actual technology and execution risk. Tier 1 is not automatically the best fit for a smaller FOAK project.
• Know where the performance guarantee comes from. If the EPC is not wrapping the risk, understand what your company will need to stand behind instead.
• If you unbundle engineering, procurement, and construction, assign clear ownership to every interface the EPC would otherwise manage.
• Consider progressive design-build or separating design from construction when you want more control without bringing the entire process in-house.
• Bring in project builders who have delivered something larger or more complex than what you are attempting.
• Let engineering maturity drive cost certainty: do not treat a Class 5 estimate like a Class 2 estimate.
• Build contingency around specific work packages and maintain a risk register that actively shrinks as those risks are retired.
• Give both the project schedule and your corporate runway more room than the base case suggests.

For a first commercial project, the ending note: Consider sacrificing first project economics to get the thing built well, and to have the opportunity to learn from working alongside project partners (developers, EPCs, lenders, CFOs) who may be “oversized” for what you’re building today, but are “fit-for-purpose” doing this repeatedly at scale. 

Thank you to Wilson Sonsini Goodrich & Rosati for sponsoring this series.

Follow our Luma calendar to register for upcoming Enduring Planet & Friends Financing Series: luma.com/enduringplanet