Transition's first loss moves to private balance sheets

KIRKBI has doubled its climate book into recycled plastics, a market the firm itself says does not work, and what it has taken in exchange is the first-loss position: the slot multilateral development banks were built to occupy, the earliest, least-proven slice of a capital stack held so that someone else can underwrite the rest. A family balance sheet standing in it instead is a better guide to who now prices the bottom of transition's stack than anything pledged from a podium this week.

The same trade appears twice more in the same stretch of days. Centrica put $13.5 million behind ERV's second fund, anchoring a vehicle that invests at Seed and Series A in electrification and thereby placing a strategic buyer inside the earliest equity of the market it buys from; Exelon's climate arm took a stake in Continuum, a vendor the company intends to deploy, which reads small as a venture bet and considerably larger as a procurement channel with equity attached.

What links the three is a shared conclusion about where option value sits in a market too early for project finance and, asset by asset, too small for the lenders chartered to bridge that gap: the first-loss layer is being repriced, and the repricing is coming from private strategic capital rather than from public institutions. This publication has argued that the structures drawing private climate money in still need a first-loss holder; this week the identity of that holder is the news.

The allocator side of the beat ran into the same wall from a different direction. Charity climate mandates are being turned into an effect test, on the observation that selling a fossil-fuel holding changes its owner rather than its emissions, with stewardship biting only at scale—the accountability half of the market discovering what the capital half already knows, that a decision about who owns an asset does not change what the asset does. Only money that builds something does, which is why the interesting moves this week were junior cheques rather than mandate rewrites.

The counterexamples are instructive precisely because of how they are shaped. Latin America's climate finance reached $108 billion in 2024 and then stopped growing, flat against the year before, while fossil fuels took $95 billion and land use outside Brazil ran 125 times short of need. Flows at that scale do not move because one family office doubles a book; they move when a state decides to make them move, and states have been deciding one project at a time.

A firm that says a sector does not work and then doubles its exposure to it is doing something coherent: buying the cheapest option in the stack, and at the early stage of a climate asset the cheapest option is the first loss. The question a position like that raises is about the buyer rather than the asset: whether a family balance sheet is underwriting to a lower return than a development bank would demand, or to the same return over a longer hold. Duration is the likely answer either way, because the capital brings no redemption cycle, no fund life forcing a sale, and no quarterly peer benchmark against which a decade-long illiquid position has to be justified.

First loss is not a euphemism for philanthropy, though public capital's defenders sometimes skip that part: it is the junior tranche, the money that absorbs the initial writedowns and therefore decides whether a project clears the return hurdle for everyone standing above it. A development bank can sit there at a concessionary price because its mandate permits it; a family office has to be paid to sit there, or believe it will be, and KIRKBI's position amounts to the second, taken in a market the firm says has not started working.

The public institutions now stand at the far end of that trade. Development finance institutions are being asked to rally private climate capital as the US backstop wavers, and the blended-finance structures doing the rallying have not stopped needing someone to absorb the first loss. The open variable is who will absorb it when the asset is a recycled-plastics plant rather than a project with a state guarantee behind it. A development bank can hold first loss on a 300 MW plant because the plant has a boundary: a site, a capacity, a completion date. It cannot hold first loss on a category.

Centrica's $13.5 million is the smaller number and the more instructive commitment, because it buys something a purchase order cannot: in a market where the constraint is the number of companies able to build and sell electrification equipment at Seed and Series A, a place on the cap table is a place in the queue, which makes the anchor commitment less a venture bet than a supply agreement financed with equity. Exelon's stake in Continuum runs the same logic one step further, taking a position in a vendor the company has already decided to deploy.

Both will get filed under corporate venture capital, which is the wrong drawer. A strategic stake in a company the parent has already chosen to buy from buys allocation in a market where allocation is the scarce good; it does not need the venture market to clear in order to pay off. Equity here is doing the job a supply contract does when supply is short, with the added advantage that the buyer can see the vendor's roadmap from the inside.

The counterweight sits in Spain, where Moeve's Onuba project needed roughly $357 million of Spanish state money to get 300 MW built at Huelva, structured across a 51/29/20 ownership split, with a 105 MW option left as the open question about whether the template continues. That is state equity doing what state equity does best, anchoring a discrete project the commercial market would not have started alone, and it is also, on this week's evidence, the ceiling of the approach: a structure that needs a government to reach first close needs an answer about the second phase, and Huelva's answer is an option worth roughly a third of the base capacity.

Climate Week's opening day made the same point with one number on each side of it: the only funded commitment was $300 million of Nigerian public money for off-grid power, against a projected 118 GW of data-centre demand that was the loudest figure in the room. One is a cheque attached to a named programme and the other is a requirement with no capital behind it, and the distance between a cheque and a requirement is where private first-loss capital has found its opening. The Nigerian commitment is the more straightforwardly useful of the two, and it is the same shape as Onuba: a state balance sheet, a named programme, a completion date. The shape is the limit.

Public money that wants to anchor a pipeline rather than a plant has to move at the speed of governance. The week's $5 billion climate pitch to three boards will clear only after three separate fiduciary reviews decide whether underwriting a pipeline that Levine says federal inaction has starved is worth the risk; KIRKBI needed one balance sheet and one conviction. The distance between those two approval processes is the entire difference.

Ares taking 80% of EDPR's contracted solar and storage portfolio shows the top of the same stack being repriced in the same week. The stake is underwritable because of the paper attached to it: 20-year offtake contracts that convert merchant exposure into a receivable. The risk does not disappear with the contracts; it settles on the offtaker's balance sheet, which the coverage flags as the exposure to watch. That is concentration of a different kind, in a different place, and likely a larger quantity of it than the first-loss cheques the week's private balance sheets wrote.

The instrument is the same one turning up across carbon removal and reforestation: Mombak closed $150 million for its second Amazon fund on the strength of a Salesforce offtake and a BNDES credit line, a combination the coverage says a reforestation fund now needs in place before it can market to LPs. Vaulted Deep's $35 million turned a carbon offtake into loan collateral, which puts lenders in the position of reading removal contracts the way they read power purchase agreements, and Symbiosis has put 40 million tons of nature-based removals behind contracts whose real content is the allocation of risk. The same purchase contract is doing two jobs: de-risking an operating portfolio for a buyer like Ares, and collateralising construction for a developer with no other asset to pledge.

Power purchase agreements have underwritten generation for decades. The range of assets now financed against a counterparty's signature has stretched to Amazon reforestation and durable carbon removal, widening the transition's construction risk across a short list of buyers whose credit is the whole underwriting file.

The two ends of the stack meet on the same point. Contracted offtake makes late-stage assets bankable, and in the process it makes the offtaker the credit; first-loss equity makes early-stage assets exist at all, and in the process it makes the holder the underwriter of last resort. Both layers were supposed to be public, and both are being taken by private balance sheets that have decided the price compensates them: a recycled-plastics plant for KIRKBI, a place in the supply queue for Centrica and Exelon, a 20-year receivable for Ares.

The second phase will tell. Huelva's 105 MW option is the clearest test available: a phase built on commercial terms would say the state-equity template outlived its first cheque, and a phase that stalls would say Onuba was a project rather than a model. The buyers now holding transition's first loss have a stake in the answer, because it decides how much of that layer ends up theirs.