Driving down the cost of green hydrogen

Start with the number that broke the hype. A kilogram of green hydrogen made in Southern Europe today runs roughly $6 unsubsidized, and that splits about evenly into the power consumed and the cost of carrying the capital that built the plant. Both halves got worse. On power, a firm green purchase agreement in West Texas that would have cost about $35 per megawatt-hour three years ago now runs $65 or more, which Garabedian attributes to surging electricity demand, most notably from data centers, and which he says is happening in cheap-power regions worldwide rather than only in the US. On capital, something genuinely strange happened: costs went up in an industry that was supposed to be descending a learning curve. Plants promised at around $1,500 per kilowatt by the large Western suppliers now land above $3,000 per kilowatt once a construction contractor has built them into a finished project. He offers two causes and is careful about the first: the aggressive forecasts were probably believed by the companies making them rather than dishonest, and simply turned out to be very hard to hit. The second is commodity inflation in the steel, titanium and nickel that electrochemical plants are made of.

Asked which half really burst the bubble, he says they are equally responsible. The asymmetry is not in the blame but in what can be done about each. He treats the power price increase as inevitable and even welcome, a consequence of an economy converting to electricity that will pull an enormous buildout of green generation behind it and eventually produce a more liquid market for renewable power and renewable molecules. Capital cost is the one he says the industry can and should work on, and if it stays where it is, the sector does not climb out of the trough.

That sets up the mechanism the episode turns on. Kann raises a claim he has heard repeated and does not trust: that a cheap enough electrolyzer lets you sidestep expensive power entirely, throttling production up and down and buying only when prices are low. Garabedian endorses it without qualification and then puts arithmetic under it. At roughly $3 per watt of capital cost, which he says is what the large incumbents offer, running at a 33% capacity factor loads so much capital charge onto each kilogram that you cannot afford to pay anything meaningful for power at all. Halve that to $1.25 or $1.50 per watt and the same strategy works, because you can buy at momentary prices below $15 per megawatt-hour and still land a cost-effective product. Kann then presses the catch: a deliberately low capacity factor means you need a great many machines to make any volume, which he says makes him somewhat more bearish near term. Garabedian’s answer moves the trade-off rather than denying it. At the $600 per kilowatt his roadmap targets, you do not need rock-bottom power at all; a 40 to 45% capacity factor with power at up to $35 per megawatt-hour gets hydrogen under $3 a kilogram. Every clause of that sentence is conditional on hitting the target.

What counts as cheap enough is then a question about geography rather than technology. Unsubsidized gray hydrogen from natural gas runs $1.50 to $2.00 a kilogram and will probably stay there while gas is plentiful. Price the carbon and the fossil molecule moves: at the carbon prices he cites for Europe and California, the equivalent cost lands between $3.50 and $4.00. So fossil parity is a $2 target in one place and a $4 target in another. He explicitly declines to build a business on that difference. Green premiums do exist today in the US and the EU and are funding real projects, but he compares them to the ones he watched in solar, where trimming the premium even slightly moved the whole market, and says he would not bet on their durability. What he will bet on is a market where green is incidental. Brazil is his worked example: a large agricultural exporter with no natural gas of its own, importing over 90% of its nitrogen with Russia as its largest supplier, paying a logistics-inflated and volatile price for it. Taking the ten-year average of about $600 a ton as the bogey, Electric Hydrogen’s analysis puts its own equipment at roughly $700 to $750 a ton of ammonia today on Brazil’s largely green grid and cheap inland power. That is not parity, but he calls it not a huge stretch from it, with cheaper-than-fossil arriving in the early 2030s if the technology roadmap and continued wind and solar cost declines both land. The buyer in that story is not paying for decarbonization. They are paying for domestic supply of a critical input, with green as the icing rather than the cake.