Onsite Oxygen Generation vs. Liquid Oxygen (LOX): CAPEX, OPEX & TCO

Oxygen is the single most expensive consumable in an intensive fish or shrimp farm after feed, and it is the one input that cannot be interrupted for even twenty minutes without losing biomass. Every RAS, raceway and intensive pond operator eventually faces the same procurement decision: install onsite oxygen generators, or contract bulk liquid oxygen (LOX) deliveries. This guide compares both routes on the metrics that actually decide the case — cost per kg of delivered oxygen, CAPEX, energy consumption, reliability, redundancy and total cost of ownership over a ten-year horizon.

FishMatch Group is independent by design: we do not manufacture, distribute or resell oxygen equipment or industrial gas, and we do not take commission from buyers. The analysis below is written to help you reach a neutral, total-cost-of-ownership decision — not to steer you toward one technology.

Executive summary

Liquid oxygen wins on low CAPEX, instant availability, high purity and simple operations, but its cost per kg is set by a supplier you do not control and it exposes the farm to delivery logistics. Onsite generation (PSA or VPSA) converts that variable OPEX into a capital asset with a predictable electricity bill, typically delivering payback in two to four years at continuous high-duty consumption — but it demands power reliability, maintenance discipline and a backup oxygen source. Most bankable projects above roughly 400–600 kg O2/day end up with a hybrid: onsite generation as the base-load supply and a LOX tank as the emergency and peak-shaving buffer.

Step 1: size your true oxygen demand before comparing anything

No comparison is meaningful until oxygen demand is quantified. Demand is driven by feed load, not by tank volume: as a planning rule, intensive systems consume roughly 0.4–0.6 kg of oxygen per kg of feed delivered, plus biofilter nitrification demand, degassing losses and transfer inefficiency. Peak demand — post-feeding, at maximum standing biomass, at the warmest water temperature of the year — is the number that sizes equipment, not the annual average.

Work through the sizing of RAS feed load to oxygen demand and dissolved oxygen management for bankable farms before requesting quotes. You can sanity-check your figures with the oxygen and biomass tools in the aquaculture calculators.

Liquid oxygen (LOX): how the economics work

LOX arrives as a cryogenic liquid in a vacuum-insulated vessel, is vaporised on site and injected through cones, diffusers or low-head oxygenators. Farm CAPEX is minimal because the tank and vaporiser are usually leased or supplied by the gas company; you pay for the concrete pad, piping, vaporiser space and injection equipment.

Cost per kg of oxygen is the whole story. Depending on region, contract volume and distance from the air separation plant, delivered LOX commonly prices between USD 0.15 and USD 0.45 per kg, with remote or low-volume sites paying materially more. On top of the molecule you pay tank rental, delivery surcharges, minimum-offtake clauses and — critically — evaporative losses from the tank, which can consume 0.3–1.5% of stored volume per day when consumption is slow.

Advantages: near-zero CAPEX, 99.5%+ purity, no electricity draw, no rotating equipment to maintain, instant full capacity on day one, and simple scaling — you just order more. Disadvantages: OPEX exposure to a supplier you cannot switch quickly, exposure to road logistics and weather, contract lock-in, and a cost line that rises with energy and freight inflation over the life of the project.

Onsite oxygen generation (PSA / VPSA): how the economics work

Pressure Swing Adsorption (PSA) generators separate oxygen from compressed ambient air using a zeolite molecular sieve, typically delivering 90–95% purity. Vacuum Pressure Swing Adsorption (VPSA) operates at lower pressure with a vacuum blower and is materially more energy-efficient at large scale, which is why most farms above roughly 1,000 kg O2/day evaluate VPSA rather than PSA.

Energy is the dominant operating cost. A well-configured PSA package including the air compressor consumes roughly 0.8–1.4 kWh per kg of oxygen produced; VPSA at scale can reach 0.4–0.6 kWh per kg. At USD 0.12/kWh, that puts the marginal production cost between roughly USD 0.06 and USD 0.17 per kg of oxygen — before maintenance, sieve replacement and amortisation. That spread against LOX pricing is where the entire business case lives.

Advantages: predictable, self-controlled cost per kg; no delivery dependency; no minimum offtake; a depreciable asset on the balance sheet; and immunity to gas-market pricing. Disadvantages: real CAPEX, lower purity (which slightly reduces transfer efficiency in oxygen cones), sensitivity to power quality and outages, compressor and dryer maintenance, sieve degradation over 5–10 years, heat rejection in the plant room, and the need for a genuine backup supply.

Head-to-head: total cost of ownership

Compare on delivered cost per kg of oxygen at the injection point over ten years, not on invoice price. A defensible TCO model includes: equipment CAPEX and installation; electricity at your actual tariff including demand charges; compressor, dryer and sieve maintenance and replacement; standby generator capacity attributable to the oxygen plant; LOX backup tank rental and delivered molecule cost; evaporative losses; tank and equipment rental; transfer efficiency differences between 93% and 99.5% purity; and the cost of downtime risk.

As a rough orientation: onsite generation rarely pays back below roughly 200–300 kg O2/day of steady demand, is usually competitive between 300 and 800 kg/day, and is strongly favoured above 800–1,000 kg/day where VPSA efficiency and volume amortisation compound. Sites with electricity above roughly USD 0.20/kWh, unstable grid supply, or seasonal demand swings often stay on LOX regardless of volume. Sites far from an air separation plant — inland Africa, Central Asia, island markets — frequently justify generation at much lower volumes because delivered LOX pricing is punitive. Run the numbers with your own tariff and quoted LOX price in the budget planner before committing.

Reliability, redundancy and the case for hybrid

Oxygen supply is a life-support system: single-point failure is unacceptable at commercial biomass density. Onsite generation introduces failure modes LOX does not have — grid outage, compressor failure, dryer failure, sieve contamination from oil carryover or moisture — so any generation-based design must carry a LOX backup tank sized for at least 24–72 hours of peak demand, automatic changeover on low pressure or low DO, and standby power for the compressor.

This is why most well-engineered farms are hybrid rather than pure: base load from generation to control cost, LOX for emergency, maintenance windows, peak feeding periods and summer temperature spikes. The hybrid also protects the LOX contract as a negotiating position: with generation installed you buy backup volumes rather than base load, which changes your commercial leverage entirely.

Procurement: how to compare quotes without being steered

Oxygen equipment and industrial gas are sold by parties with a direct interest in your choice: generator OEMs quote against LOX, and gas companies quote against generators. To stay neutral, issue one specification to both routes and require answers to the same questions. Ask each vendor for: guaranteed kg O2/day at your peak water temperature and altitude; specific energy consumption in kWh/kg at 50% and 100% load; purity and its effect on transfer efficiency in your injection equipment; sieve and compressor service intervals and replacement cost; noise and heat rejection; warranty and local service response time; and — for gas suppliers — the delivered price per kg, tank rental, minimum offtake, price escalation formula and contract exit terms.

Normalise every quote to delivered cost per kg of oxygen at the injection point, then rank. This is the same buyer-first discipline described in vendor-neutral aquaculture procurement and in our guide to choosing aquaculture equipment suppliers. If you want the comparison run independently, request a confidential RFQ — FishMatch Group is free for buyers and takes no fee from you.

Financing and bankability implications

The choice changes your financial statements as much as your plant room. LOX keeps the balance sheet light and the CAPEX request small, which can help a first-round debt package clear — but it leaves an uncontrolled OPEX line that lenders stress-test aggressively in downside scenarios. Onsite generation raises the capital request and adds electrical infrastructure, but produces a defensible, modellable cost per kg that improves EBITDA predictability over the loan tenor.

Lenders will ask which route you chose and why, and they will expect the TCO model, the redundancy design and the backup contract in the technical package — see aquaculture project financing: what lenders require. If the oxygen plant is part of a larger capital raise, the financing center outlines how equipment CAPEX is typically structured.

Decision checklist

Choose LOX when: peak demand is below roughly 300 kg/day; the grid is unreliable or electricity exceeds roughly USD 0.20/kWh; you are within economical delivery distance of an air separation plant; CAPEX is constrained; or demand is seasonal and highly variable.

Choose onsite generation when: demand is continuous and above roughly 400–600 kg/day; electricity is reliable and reasonably priced; delivered LOX pricing is high due to distance or low competition; you have or can hire the maintenance capability; and you can fund the CAPEX and a LOX backup.

Choose hybrid — the default for most commercial projects — when supply security is non-negotiable and you want cost control on base load with contractual and physical resilience on top.

Related reading

Continue with dissolved oxygen management for bankable farms, sizing a RAS feed load to oxygen demand, energy efficiency in RAS and pond systems and aeration in tropical shrimp ponds. Pillar guide: RAS vs. Sea Cage Farming. Full outline: RAS Planning Topic Cluster.