Aluminium as an off-grid firming fuel

I did two nerdy things last week and thought I'd bring them together here.

First, I published a chart comparing capex and opex for firming power options for an off-grid data centre in the Mid West of Western Australia — a 100MW flat load near Geraldton, with wind, solar and batteries optimised at every point, every megawatt hour served and zero fossil fuel. Second, I listened to an interesting episode of Latitude Media's Catalyst podcast, with Shayle Kann interviewing Voya Energy co-founder and CEO Richard Wang on using aluminium pellets as an off-grid fuel.

Voya's system is an electrochemical generator rather than an engine — aluminium pellets react with air and water in a metal-air cell, leaving aluminium trihydrate powder as the residue. No combustion and no exhaust, and a fuel that stores in a container indefinitely, which is a useful property when the job is covering the occasional bad fortnight of weather.

So I added it to the chart. The four green dots compare the aluminium fuel option to other forms of low-carbon firming energy. The key comparison metric is average system levelised cost of electricity, LCOE — the average of everything over a year, not just the firming cost.

Voya don't publish information on their capex, so I've assumed two bookends from CSIRO's GenCost 2025-26: at the low end, the same capital cost as the reciprocating engine on the chart (about AUD$2.4m per MW installed); at the high end, the same cost as a fuel cell (about AUD$6.9m per MW), which is arguably the closer cousin given the technology is electrochemical.

Richard explained on the podcast that Voya are targeting scrap aluminium as the fuel. This chart tells you why. I've used low-grade scrap at AUD$1.20/kg as the low end fuel cost — Perth yards currently pay AUD$0.80 to AUD$2.50/kg — and virgin aluminium at the LME price, roughly AUD$5.00/kg, as the high end. At Voya's claimed 57% conversion efficiency, scrap works out at about AUD$245/MWh as a running cost; virgin metal is north of a dollar per kilowatt hour, and the optimiser dispatches it accordingly.

With the most optimistic mix (engine-like capital cost and scrap aluminium as the fuel) costs could be similar to burning biomethane in a gas engine: AUD$137/MWh for firmed annual power, within a dollar of the cheapest option on the chart.

At the other extreme (fuel cell capital cost and virgin aluminium) it's by some margin the most expensive firming technology on the chart at AUD$199/MWh. But it's still cheaper than trying to get to 100% carbon-free on renewables and batteries alone, which is AUD$230/MWh in my model — the last few percent of unfirmed reliability is bought with a lot of idle capacity.

Two caveats worth stating plainly. The 57% efficiency is Voya's own figure, from their Series A announcement — there's no independent test data yet, though it is consistent with their published storage-density claims. If the real number were 45%, the best case moves from AUD$137 to the low AUD$140s. And AUD$1.20/kg is what a yard pays for raw scrap — a pelletised, quality-controlled fuel will cost more, though the trihydrate residue has resale value. In the mid case the model would consume around 20,000 tonnes of scrap a year, so fuel supply at that price is a real question, not a detail.

My read: aluminium fuel looks like a credible fourth option alongside biomethane, renewable diesel and hydrogen for off-grid firming, with better storability than any of them. Whether it wins outright depends on where the capex lands between my two bookends, and nobody outside Voya knows that yet.

All calcs were performed in our techno-economics platform, Levelised. Drop me a line if you'd like to play with the model or learn more.