← All perspectives

Company · Sustainable Energy · Equity · 2021–2023

The Cost-Curve Test: Why We Built a Hydrogen Platform, Then Walked Away

We saw hydrogen as a candidate secular winner and did the work to own it — assembling a three-company platform across production, storage, and demand. Then we ran it through the one test the theme could not pass: cost. This is why we developed the strategy, and why we correctly chose not to fund it.

Being early to a secular winner is worth a fortune. Being early to one that never arrives is worth nothing — and at the start, the two look identical.

Our mandate is to find the next structural winner before it becomes consensus, and to build a position while the field is still open. Most of what we develop, we fund. Some of it we don't — and the discipline to tell those apart is as much a part of the record as any deal we close. A serious investor's anti-portfolio, the good ideas correctly declined, is as revealing as the trophy case and usually more honest about how the decisions were actually made. Hydrogen was the hardest version of that decision we have made: a theme we believed in, backed by a platform we assembled ourselves, that we ultimately declined to capitalize. We think we got it right, and the market has since agreed.

The Mandate: Own the Winner Before the Crowd

The case for hydrogen was, and remains, real. Electrification handles most of the energy transition — but not all of it. There is a stubborn tier of the economy that electrons cannot easily reach: steelmaking, ammonia and fertilizer, refining, shipping, heavy long-haul transport, and long-duration storage for a grid running on intermittent wind and sun. Green hydrogen is one of the few molecules that can decarbonize that tier directly. It is the transition's designated tool for the jobs electricity can't do.

That is a genuine secular thesis, and by the early 2020s it had policy behind it. The Inflation Reduction Act's 45V credit offered up to $3.00/kg for clean hydrogen — one of the most generous production subsidies in the entire energy transition. Governments set formal cost targets. Capital flooded in. For an investor whose job is to be early to structural change, hydrogen checked the first box that matters: a large, policy-backed problem that only this technology could solve.

We did not want exposure to the theme. We wanted to own a piece of the value chain.

What We Built: A Platform, Not a Bet

The 1V1sion approach to a secular theme is not to buy a stock and hope. It is to build the asset. So we married the macro thesis to M&A and development, and assembled a three-company platform spanning the hydrogen value chain end to end:

  • Production — an electrolyzer intellectual-property play sourced from a leading American university lab, the technology that splits water into hydrogen and the cost center that decides whether green hydrogen is ever competitive.
  • Storage — a battery and energy-storage technology company, the firming layer that turns intermittent renewable power into the steady input an electrolyzer needs, and stores what the system produces.
  • Demand — an aerial "drone power plant," a mobile generation platform capable of running on hydrogen, giving the chain a differentiated, high-value application rather than betting solely on commodity offtake.

The logic was vertical integration of an emerging industry: control the production technology, own the firming that feeds it, and anchor a demand application that pays a premium. Assembled by active capital, this was a real platform, not a paper thesis — proprietary IP, an operating storage technology, and a use case with its own economics.

Every piece was defensible on its own. The question was never whether the pieces were good. It was whether the system they served could clear a price.

The Cost-Curve Test

Here is the discipline that governs everything we do with a new technology, and the reason we passed. We call it the Cost-Curve Test: a secular theme is only investable when its cost curve is bending toward competitiveness on a timeline you can actually underwrite. Promise is not a price. Enthusiasm is not a learning rate. A technology can be inevitable in the long run and uninvestable in the run your capital lives in.

Hydrogen failed that test, and the numbers were not close. Green hydrogen was producing at roughly $4–8/kg — as much as $12/kg delivered — against grey hydrogen from natural gas at about $1–2/kg. The single most credible target in the field, the U.S. Department of Energy's "Hydrogen Shot," aimed at $1/kg — but not until 2031, with an interim $2/kg pencilled for 2026 that the industry is now missing. Even a $3/kg subsidy did not close a gap that ran three-to-six times the incumbent cost.

It is worth being precise about why hydrogen is harder than the transition's success stories. Solar modules and lithium-ion batteries each rode a steep experience curve — cost falling roughly 20% with every doubling of cumulative production, compounding into declines near 90% over the 2010s. Those are manufactured objects; scale makes them cheap. Green hydrogen has no single such curve, because its delivered cost is dominated by two things scale does not fix: the price of the renewable electricity it consumes, and the energy lost making, storing, and reconverting the molecule. You can mass-produce the electrolyzer. You cannot mass-produce cheaper physics. That is why solar and batteries crossed into the money while hydrogen kept promising to.

And the gap was structural, not just early. Electrolysis carries an unavoidable thermodynamic penalty — you spend more energy making and later using hydrogen than the hydrogen returns — so the green molecule is chained to the price of renewable power and the capital cost of the electrolyzer, with scarce iridium and platinum in the stack. The curve was bending, but slowly, and every credible forecast kept sliding the crossover point to the right. We could build the platform. We could not underwrite the cost curve reaching parity on any schedule our capital could accept. After a long development effort, we pulled the plug.

We Were Right — the Market Caught Up

Passing on a theme everyone is chasing is uncomfortable. The validation came fast.

By 2024 and 2025 the hydrogen build-out entered what the industry now calls its viability gap — a wave of cancellations as the cost reality met the announcements. Nel scrapped a planned U.S. electrolyzer factory. Topsoe paused its flagship American plant. CF Industries cancelled its green-hydrogen project. Cummins stopped pursuing new electrolyzer activity after roughly $458 million in related charges, its management stating that demand had "dried up." The projects died on precisely the problem the Cost-Curve Test flagged: the cost to produce ran far above what any buyer would pay, and the bankable offtake never materialized.

The subsidy did not save them, and that too was foreseeable. A $3/kg credit only matters if a buyer signs a long-term contract at a price that clears — and buyers would not, because the delivered molecule still cost multiples of the grey hydrogen already plumbed into their plants. Policy can subsidize supply; it cannot manufacture demand at an uncompetitive price.

We reached that conclusion before the retrenchment, not after it. That is the difference between discipline and hindsight.

What Passing Protected

The Sunk-Cost Antidote is the discipline of counting a good pass as a win — the refusal to fund a thesis simply because we had already spent time on it. Hydrogen is its clearest case.

  • Capital preserved. The single most valuable thing an early investor can do in a theme priced for perfection is not lose money in it. Every dollar we did not commit to a cost curve that never bent is a dollar that compounded elsewhere.
  • The thesis, not the timing, was right — and we separated the two. Hydrogen may still win its designated jobs, later, when renewable power is cheaper and the stack is cheaper still. Being right about the destination is not the same as being right about the entry, and marrying the two is how investors get trapped in a good idea at a bad price.
  • The platform work was not wasted. Sourcing an IP play, an energy-storage technology, and a novel power application taught us the value chain cold. That knowledge is an option we still hold; the check is one we chose not to write.

We share the discipline here, not a trade. We built a real hydrogen platform because we believed in the destination, and we declined to fund it because the cost curve would not cooperate on a timeline we could underwrite. The market's 2024–2025 correction settled the question. Being early is only an edge when the curve is bending fast enough to meet you — and the principal's job is to know the difference before the capital goes in, not after.

This is how we underwrite a new technology: destination and price are separate questions, and we will pass on a theme we believe in when the cost curve won't clear on a schedule our capital can accept. If the energy transition is on your radar, we're glad to walk through the work — including the parts we chose not to fund.