The conclusion first
After six years and roughly $180,000 in tracked industrial gas spending, here's my answer to the question I get asked most: if you're sourcing gas for a plant in the Cleveland area, the supplier with the lowest per-unit quote is usually the one that will cost you the most by year-end — by 12–24%.
That holds for oxygen, nitrogen, argon, and welding mixes. It holds whether you're benchmarking a regional distributor against a national player like Air Products, or two regionals against each other. The reason isn't dishonesty. It's that per-unit pricing is the only number everyone publishes, so it's the only number most buyers compare.
Three things I'd want a buyer to take away:
- Cylinder rental, delivery minimums, and expedite fees absorb 15–25% of a typical annual gas budget. None of it appears on the quote sheet.
- For high-volume users, on-site generation — PSA nitrogen, membrane separation — changes the math entirely. The premium is real, and it's frequently worth it.
- Hydrogen is what everyone wants to discuss in 2025, and almost nobody in this region has the volume to justify it yet. More below.
What follows is how I got there, what it cost me to learn, and the cases where this logic breaks down.
Why weight this (and where I'm just a guy with a spreadsheet)
I'm procurement manager at a 210-person metal fabrication shop about 25 minutes outside Cleveland. I've owned our industrial gas budget — roughly $180,000 a year across bulk and cylinder supply — for six years, negotiated with 11 vendors, and logged every invoice in our cost tracking system since Q1 2019.
I'm not a gas engineer. If your question is about tank sizing, manifold design, or whether a specific purity grade meets your process spec, that's outside my lane — talk to your process engineer or the supplier's technical rep. What I can speak to is what shows up on the invoice, and how often the invoice doesn't match what you were quoted.
The $8,400 mistake, in detail
When I took over gas procurement in 2019, I assumed the job was three quotes and the lowest number wins. That's what the old spreadsheet did, and it seemed reasonable at the time.
In 2020 we ran a full re-bid. Vendor A quoted $61,400 for the year, all-in, including monthly service and cylinder management. Vendor B quoted $56,900. A $4,500 spread — and I nearly signed with B on the spot.
What stopped me was a line buried in the appendix: $65 per month, per manifold, for cylinder rental. We run 14 manifolds. That's $10,920 a year, and it wasn't in the headline figure. Add $180 per expedited delivery (we averaged four a month through the busy season, so call it $8,600), plus a $22 hazmat surcharge on every cylinder swap.
Vendor B's real annual cost landed around $69,800. Vendor A's $61,400 included cylinder management and carried no expedite fee up to six deliveries a month. We went with A.
The "cheaper" quote was 13.7% more expensive. I built a TCO calculator that week and haven't run a renewal without it since.
That's the shape of the problem — not malice, just a different fee structure that the summary sheet conveniently omits (note to self: always request the fee schedule as a standalone document, never as a paragraph inside the proposal).
Where the money actually goes
Across six years, about 18% of our gas spend has fallen outside the per-unit rate. The usual suspects:
- Cylinder demurrage. Rentals accrue monthly whether you draw the gas down or not. If your usage is seasonal, you're paying rent on idle steel.
- Delivery minimums. A $150 drop-off fee on a $400 order is a 37% markup dressed up as logistics.
- Product loss on bulk liquid. Liquid nitrogen and oxygen vent off. A poorly sized tank loses 3–5% a month. That's your gas, evaporating, and you're already invoiced for it.
- Return handling. Some contracts charge cylinder inspection on return. Ours did, until we negotiated it out in 2022.
None of these are scandalous. They're simply not on the quote sheet, and if you don't ask, nobody volunteers them.
The part where the industry moved under my feet
What worked in 2019 is not the playbook for 2025. Two things changed.
First, on-site generation got genuinely competitive for mid-size users. PSA nitrogen systems and membrane separators — Air Products' PRISM units are the reference point most vendors benchmark against — now pencil out for shops above roughly 8,000–10,000 standard cubic feet per hour. Five years ago that threshold sat considerably higher. The trade-off is capital outlay and maintenance you now own, so it isn't free. But the crossover point has moved, and if you're still buying all your nitrogen as liquid at high volume, it's worth rerunning the numbers.
Second, hydrogen. Air Products has pushed hydrogen hard, and it's a real operating business for them, not a marketing exercise. But here's what nobody says out loud: most industrial buyers in Northeast Ohio don't have the volume to justify hydrogen supply in 2025. The purity specs are well defined (ISO 14687 is the relevant standard for fuel-grade hydrogen), the delivery infrastructure exists, and the cost floor is still high.
Be careful with sustainability claims on both sides of the table. Per the FTC Green Guides (16 CFR Part 260), environmental claims like "green hydrogen" require substantiation, and "carbon-free" carries a specific definition. If a supplier can't walk you through the production pathway — electrolysis, SMR with capture, byproduct — get it in writing before it goes anywhere near your ESG report. I've watched procurement teams paste supplier marketing language straight into a sustainability filing. That's a bad afternoon waiting to happen.
Where this doesn't apply
Three honest exceptions.
If you run fewer than about ten cylinders a month, ignore most of this. Your absolute spend is too small for TCO analysis to beat the hourly rate of whoever is doing the analysis. Buy from the supplier that delivers reliably and answers the phone.
If you need a specialty gas with a single qualified source — calibration mixes, ultra-high-purity electronics grades — you don't have a comparison problem, you have a qualification problem. Switching costs dominate everything else on this page.
And if you're in a genuine supply crunch, pay the expedite fee and move on. I've watched a plant manager spend three days negotiating a $600 surcharge while a line sat idle. The arithmetic doesn't favor you there.
One last gap I'll admit to: I don't have hard data on industry-wide delivery reliability, only our own six-year record. Anecdotally, we've had two late bulk deliveries in six years with our current supplier and eleven with the one before. Two-versus-eleven isn't a statistically rigorous sample. But I know which contract I'm renewing.