Hydrogen blending: what today's equipment actually accepts

Somewhere between the brochure that promises a hydrogen-ready future and the cynic who calls it all marketing sits a narrow set of facts an owner can actually use. This paper collects them.

Section 01Two different questions wearing one word

When "hydrogen" appears in an on-site power conversation, it usually means one of two very different things, and conflating them is the first error to avoid.

The first is pipeline blending: hydrogen injected into the common natural gas system, arriving at your meter mixed into the fuel everyone receives. This is a system-level question governed by utilities and regulators, not by anything you buy. The second is behind-the-meter blending: hydrogen you procure separately, delivered or produced on site, blended into the fuel stream of your own generation equipment. This one is a project decision, and it is the one vendors are usually selling when they print "hydrogen-ready" on a data sheet.

The two questions have different economics, different constraints, and different timelines. An honest screen treats them separately.

Section 02What the system-level studies actually found

California examined pipeline injection directly. In 2022 the California Public Utilities Commission released an independent study on injecting hydrogen into the state's natural gas systems, and its findings were measured: blending appears feasible at low concentrations, while higher concentrations raise material questions about pipeline materials, end-use equipment, leakage, and safety that require further work before system-wide adoption.¹ There is still no standardized statewide blend percentage for the common system, and gas utilities' tariff rules governing gas quality, including PG&E's Rule 21 for gas, define what the system will accept.²

The practical consequence for an owner is simple: do not build a project economics case on the assumption that the gas arriving at your meter will decarbonize on a schedule. If cleaner pipeline gas arrives during your project's life, it improves your emissions math for free. Nothing about your equipment selection should depend on it.

Section 03Equipment tolerance, class by class

Behind the meter, tolerance for hydrogen varies by equipment class, and the honest generalization is that every class has one: a stated blend limit beyond which warranty, performance, or safety margins change. The classes differ in character.

Note what this table does not say: it does not say any class "cannot" run hydrogen, and it does not say any class solves your problem because it can. Tolerance is a property of a specific model, a specific blend, and a specific warranty document. Class-level generalizations, including ours, are only the map to the questions.³

"Hydrogen-ready" is not a specification. A blend percentage, on a named model, in a warranty you can read, is a specification.

Section 04The supply question nobody prices

Equipment tolerance is the smaller half of the problem. The larger half is supply: where the hydrogen physically comes from, at what delivered cost, with what reliability.

Today, an owner buying hydrogen for blending faces delivered costs that are a multiple of natural gas on an energy basis, a logistics chain (tube trailers, on-site storage, or on-site electrolysis) that adds capital and permitting scope of its own, and a young market whose pricing depends heavily on federal credit policy still being implemented through rulemaking.⁴ None of that makes hydrogen wrong. It makes hydrogen a separately priced project, with its own line items, rather than a checkbox on a generation purchase.

The disciplined way to represent this in a study is a sensitivity case, clearly labeled illustrative: what happens to lifecycle cost if a defined blend at a defined delivered price is introduced in a defined year. If the project only pencils in that case, the project does not pencil.

The blend arithmetic, run once, honestly

Most hydrogen conversations improve immediately when one piece of physics is put on the table: hydrogen carries roughly one third the energy of natural gas per unit of volume. The consequence, using deliberately illustrative round numbers, is that a blend of 20 percent hydrogen by volume delivers only on the order of 7 percent of the fuel's energy from hydrogen. Volume percentages are what blend specifications quote; energy percentages are what your economics and your emissions actually follow. Conflating the two overstates every benefit by roughly a factor of three.

Run the emissions line with that correction and the sober picture emerges: a 20-percent-by-volume blend reduces combustion carbon dioxide by roughly the hydrogen's energy share, several percent, not twenty. Whether even that reduction is real at the system level then depends entirely on how the hydrogen was produced, a question the federal clean-hydrogen credit rules devote their full complexity to precisely because production routes differ so widely in their own emissions.⁴ A paper exercise: if the hydrogen displaces one million British thermal units of gas priced at a few dollars, and the delivered hydrogen supplying that energy costs a multiple of that figure, the implied cost per ton of carbon avoided lands far above what the same dollars would abate elsewhere at most plausible near-term hydrogen prices. Your numbers will differ; the discipline of computing them, labeled illustrative, before signing anything, will not.

The same arithmetic clarifies what blend capability is actually worth today: it is a hedge against a future in which delivered hydrogen becomes cheap where you operate. That future may arrive; regional infrastructure programs are working toward it. But a hedge has a fair price, and the fair price of an option on an uncertain fuel is modest. Pay modest prices for it, demand the tolerance in writing, and let the machine you buy stand on the fuel you can price this year.

Section 05What to pay for now, and what to defer

Pay for now:

• Optionality that is cheap. If two otherwise comparable machines differ in stated blend capability, and the more capable one costs little or nothing extra, take the capability. Free options are worth holding.
• Written tolerances. Whatever you buy, obtain the manufacturer's blend statement for your model in writing, including warranty treatment. This costs a request.
• Space and routing provisions. If your site plausibly hosts future hydrogen logistics, reserving the pad space and pipe routing in the initial design is inexpensive; retrofitting it is not.

Defer, until the numbers exist:

• Premiums for capability you cannot fuel. Paying materially more for high-blend capability, with no priced hydrogen supply on your horizon, is buying insurance against a risk you have not quantified.
• Economics that depend on future fuel switching. Any model whose recommendation flips based on assumed future hydrogen prices is telling you about the assumption, not the site.
• On-site production, absent a separate business case. Electrolysis on site is its own project with its own load, water, permitting, and economics. It deserves its own analysis, not a line item.

Section 06The honest screen

Reduced to a checklist, the hydrogen screen inside a technology-neutral study asks five things: What blend does the specific model accept, in writing, with what warranty treatment? What does the blend do to emissions and permitting at your site? Where would the hydrogen physically come from, and at what delivered cost per unit of energy, today? What does the sensitivity case show, labeled illustrative? And would the recommendation survive deleting hydrogen from the analysis entirely?

That last question is the one that keeps everyone honest. A project that stands on today's fuels, with cheap optionality for tomorrow's, is robust. A project that needs tomorrow's fuel to justify today's purchase order is a bet, and bets belong in the memo's risk section, not its conclusion.