00Executive summary
One molecule, a rainbow of recipes — and only two numbers that matter.
"Hydrogen is the future" is a slogan, not an answer. Which hydrogen? The colors are a marketing vocabulary for the production pathway, and they hide a 3–6× cost spread and a 40× carbon spread. Two numbers settle every argument: dollars per kilogram, and kilograms of CO₂ per kilogram of hydrogen. Map all the colors on those two axes, overlay the subsidies, and the ranking falls out — with one twist that surprises almost everyone.
Here is the twist. Honest green hydrogen costs ~$4.5/kg and loses to grey on raw price — until the 45V credit, earned only by clearing that carbon line, drags its net cost down to ~$1.50/kg, level with the dirtiest incumbent. The chemistry didn't change; the subsidy did. That is the whole episode: the 4 kg line, not the color, gates the money. And in 2025 a budget bill moved that line's deadline — a rule change that re-scores every project at once.[5]
What's inside
01The colors, fast
Same H₂, different recipe — and the catch in each
Grey — steam-methane reforming of natural gas, no capture. The incumbent: ~$1–2/kg, ~9–12 kg CO₂/kg, and ~95% of the world's ~97 Mt of hydrogen in 2023 was made from unabated fossil fuels.[1][7][9] The bar everything else has to beat — and the thing being greenwashed whenever someone calls hydrogen "clean" by default.
Brown / black — coal gasification. Even dirtier than grey (roughly ~18–20 kg CO₂/kg, route-dependent), and dominant in China's installed base.[7] The incumbent to retire.
Blue — grey plus carbon capture (SMR or ATR + CCS), ~$2–3/kg.[9] The honesty test has two parts, both cited: the real capture rate (proponents need "well over 90%"; real-world SMR plants capture far less) and upstream methane leakage (Howarth & Jacobson modeled 1–4%, ~3.5% midpoint, and found blue's lifecycle footprint "quite high" — in the worst case higher than just burning the gas).[8] Real bet only if it's honest about both; a fossil lifeline otherwise.
Turquoise — methane pyrolysis: CH₄ → solid carbon + 2 H₂, so the carbon comes out as a solid and there's no CO₂ to capture or store.[11] Monolith's Olive Creek 1 (Lincoln, Nebraska) has run at commercial scale since 2020 (~2,500 t/yr H₂ + ~14,000 t/yr carbon black).[10] Base-case economics reach ~$1/kg — but only if the carbon-black byproduct sells, and flooding that market would crash its price.[11] Watch: byproduct economics are the whole ballgame.
Green — electrolysis from renewables. The headline: ~$3–6/kg today (IEA's band stretches to $4–12, falling toward $2–9 by 2030), near-zero operational carbon if the power is genuinely clean.[2][9] The catch is in that "if": on a dirty grid, electrolytic hydrogen can be dirtier than grey — which is exactly why 45V scores lifecycle carbon, not nameplate. And electricity is 55–70% of green's cost, so green's price is really the power price in disguise.[15][3]
Pink (nuclear electrolysis) — same near-zero carbon as green; its edge is a high capacity factor (an electrolyzer runs ~90%+ of the time on nuclear vs ~30–50% on solar/wind), which slashes per-kg capital cost. Qualifies for 45V like green. We ship no standalone $/kg — kept qualitative.[3] Yellow (solar-only electrolysis) folds into green. White / gold (naturally occurring geologic H₂) is the wildcard: the Bourakébougou, Mali reservoir powers a village and was characterized in Nature Scientific Reports as spontaneously recharging, with newer claims in France's Lorraine basin and elsewhere.[12] "Cheap if real" — but at-scale extraction economics are unproven, so we ship no cost number. Huge if true.
02The comparison matrix
Every color on cost, carbon, maturity — and the verdict
| Color | How made | Cost $/kg | CO₂ kg/kg | Maturity | Where it pencils | Verdict |
|---|---|---|---|---|---|---|
| Grey | SMR, no capture | $1–2 [9] | 9–12 [7] | Commercial (95%+ supply) [1] | Today's industry, until carbon is priced | Incumbent |
| Brown/Black | Coal gasification | ~$1–2 | ~18–20 [7]·MED | Commercial (China base) | Nowhere clean | Overhyped |
| Blue | SMR/ATR + CCS | $2–3 [9] | ~1 → >grey [8] | Components commercial | High capture + low leak only | Real if honest |
| Turquoise | CH₄ pyrolysis → H₂ + solid C | ~$1* [11] | low if heat clean [11] | Commercial demo (2020) [10] | When carbon black sells | Watch |
| Green | Renewable electrolysis | $3–6 (–12) [2][9] | ~0 if clean power [3] | Commercial, scaling | Subsidy + cheap-power niches | Real Bet |
| Pink | Nuclear electrolysis | power-driven | ~0 [3] | Rides nuclear build-out | High-capacity-factor sites | Watch |
| White | Geologic, extracted | unproven [12] | ~0 (natural) | Exploration only | If/where it's found | Wildcard |
*Turquoise ~$1/kg holds only if the carbon-black byproduct sells. Point values are illustrative midpoints of the sourced ranges; brown's exact carbon intensity is a medium-confidence peer-LCA figure (lead with grey's sourced 9–12 and read brown as "even dirtier"). Pink and white carry no cost number on purpose — the sourcing gate forbids a $/kg we can't stand behind.
03The cost-vs-carbon map
The money chart — and the subsidy zone that decides who gets paid
04The money map
Incentives are tiered by carbon — so policy, not chemistry, reshuffles the board
Every major subsidy is keyed to lifecycle carbon intensity. A color's kg CO₂/kg decides which credit it qualifies for, and the credit decides whether it pencils. Five levers run the table:
| Lever | What it pays | Who it favors |
|---|---|---|
| US 45V Clean H₂ PTC | Up to $3/kg, four-tier by lifecycle CI. Full credit at ≤0.45 kg; zero above 4 kg. Final rules (Jan 2025) lock the "three pillars" — additionality, hourly matching by 2030, deliverability.[3][4] | Honest green & pink (clean power) |
| US 45Q CCS credit | $85/t CO₂ for saline storage, $180/t for direct air capture (2024–26 vintages).[6] | Blue & any capture play |
| Canada Clean H₂ ITC | A refundable 15–40% investment tax credit on project capital, tiered by the same lifecycle carbon math — full 40% under 0.75 kg, nothing at 4 kg or more. Methane pyrolysis (turquoise) added Dec 2024; capture equipment gets its own 50% CCUS ITC.[16][18] | The same 4 kg gate, paid as capex |
| Canada carbon price | The industrial price trajectory was redrawn May 2026: from a line pointing at $170/t by 2030 down to ~$115/t ($95 held in 2026; $130 by 2035).[17] | Re-prices every capture project |
| EU Hydrogen Bank | A fixed €/kg premium over 10 years. First auction (2024) cleared at €0.48/kg (ceiling €4.5); awards ran €0.33–1.88/kg; a later round cleared near €0.60/kg.[13] | RFNBO (EU-additional) green |
| EU CBAM | The carbon border tariff now lists hydrogen as a covered good; definitive charging phases in from 2026, pricing carbon into imported grey/blue.[13] | Penalizes dirty imports |
| China | Green-H₂/ammonia standards + provincial mandates — but the real lever is the electrolyzer cost lead: Chinese alkaline systems run ~¼ the Western price (~$185–200/kW in 2024).[14] | Green capex, globally |
05The flip: what 45V does to green
Green loses on raw cost — and wins on net cost, because of one credit
This is the chart the slogan never shows you. On raw cost, honest green (~$4.5/kg) is 2–3× the grey incumbent. Apply the 45V credit it earns by clearing the carbon line, and its net cost collapses to ~$1.50/kg — into the grey band. Nothing about the molecule changed.
Green hydrogen cost — before vs after the $3/kg 45V credit
The credit acts as a ~$3/kg vertical shift. Blue makes a smaller move (it must first clear the 4 kg gate at all, then earns a far smaller tier); grey and brown earn nothing, because they never clear the gate. Run your own numbers in the companion model. [3][4][8][9]
The companion workbook (Bankable-RB02-Hydrogen-Colors-TEA.xlsx) makes this editable: change any color's cost or carbon intensity and the 45V credit, the net-cost flip, and the carbon-price breakeven all recompute live. It also answers the carbon-price question directly — with no subsidy at all, green beats grey on total cost once carbon is priced at roughly $300/t; for context the EU ETS has traded ~€60–90/t and 45Q already pays $85/t for stored CO₂.[6]
06The ranking & per-color verdicts
Two real bets, two to watch, one wildcard, two to retire
Green — Real Bet, in its niches and with subsidy + cheap power. It is the headline for a reason: a clear carbon advantage, a credit built for it, and a global capex curve bending down (China). The honest caveat is the "if the power is clean" pillar and the power-price sensitivity. Blue — Real Bet only if it's honest about capture rate and methane leakage; a fossil lifeline otherwise, and the one most likely to be oversold. (We red-team it in a later episode.)
Turquoise — Watch: real if the carbon-black market holds, with a genuine commercial-scale plant already running. Pink — Watch: rides the nuclear build-out; its capacity-factor edge is real but it's hostage to new nuclear. White — Wildcard: enormous if the geology and extraction economics prove out; unproven at scale today, so a watch-item, not a bet.
Grey & brown — the incumbents to beat, overhyped only when sold as "clean." Grey is the honest, cheap, dirty baseline every clean color is measured against; brown is the one to retire first.
The through-line: the color wheel is marketing; the carbon line and the credit schedule are the business. Which color we tear down first is up to you — the poll on the episode orders the deep-dives.
07Sources
18 sources — agency / government / peer-reviewed lead. Original-language titles preserved. Confidence flags carried through from the research dossier.
Want the editable model and the deep-dives?
This report is the free tier. The numbers here are live in the companion workbook, and each color gets its own teardown next.
Each color gets its own full teardown — whichever wins the episode poll goes first. Cast your vote on the episode. · bankable.show
08Method & disclaimer
Every quantitative claim is traced to a source above. Point values used in the matrix, the scatter, and the companion model are illustrative midpoints of the sourced ranges (e.g. grey $1–2/kg → $1.50), chosen to demonstrate the 45V mechanics — not project-grade cost estimates. Confidence flags are carried through: brown's carbon intensity is medium-confidence; white hydrogen's cost is low-confidence and therefore never shipped as a $/kg; pink is kept qualitative. The 45V credit schedule and the three pillars are as of the January 2025 Treasury final rules, with the 2025 OBBBA construction-deadline change noted. The companion workbook is illustrative and pending qualified-reviewer (P.Eng) sign-off before any external use. Numbers sourced 2026-06-17; the Canada policy layer (Clean Hydrogen ITC, CCUS ITC, the May-2026 carbon-price revision) added 2026-07-02.