Concept

Feedstock & energy cost

Overview

Feedstock and energy cost is the recurring cost of the raw materials and energy a process consumes — how much it uses per unit of product (from the mass and energy balance) × the price of each input. For most commodity processes it’s the largest part of variable opex and a dominant share of total production cost.

Body

What it is. The purchased inputs crossing the system boundary inward — raw-material feedstocks and energy (electricity, fuel gas, steam) — priced and summed:

feedstock & energy cost per unit = Σ (consumption per unit of product × input price)

It has a consumption side and a price side, and the cost is their product.

Consumption is physics. Feedstock per unit comes from the mass balance through conversion and yield; energy per unit is the energy intensity from the energy balance. These are properties of the process and its efficiency — better conversion and tighter heat integration lower consumption.

Price is markets. Input prices (electricity per MWh, gas per GJ, an intermediate per tonne) are set by markets, vary by location and time, and are often volatile. Not properties of the process at all — which is why they’re the natural axis of a one-way sensitivity and the variable that distinguishes scenarios.

Why it dominates. For a commodity made at scale, feedstock and energy is commonly the majority of variable opex and a large fraction of levelized cost — the product sells near commodity margins, and converting cheap inputs into product is the business. Where the system boundary is drawn decides which input price governs: buy hydrogen at the fence and the H₂ price dominates; make it inside from electricity and the power price does.

Limits & typical error

• A product of two uncertain factors, so both errors compound — cost = consumption × price; a 15% error in either moves cost ~15% before the other is counted.
• Price volatility usually dwarfs the engineering uncertainty — energy and feedstock prices swing more, and faster, than a well-estimated consumption figure, so the input price is frequently the single largest uncertainty in the whole cost and the dominant sensitivity driver. A point estimate on one day’s price hides that range.
• The governing price depends on the boundary — reading a cost as feedstock-price-driven when the boundary actually encloses the feedstock’s production (making it energy-price-driven) misattributes the dominant input; fix the system boundary first.
• Average prices misrepresent time-coupled operation — a flat annual price applied to a plant that deliberately runs when input is cheap (an intermittent green plant chasing low-priced power) overstates its energy cost; capacity factor and realized price aren’t independent.
• Embedded energy can be double-counted or dropped — energy embodied in a purchased feedstock (the power in bought hydrogen) vs. energy bought directly must be counted once.

Mini-example

Green ammonia’s energy cost is mostly one line. A tonne of NH₃ takes ~10 MWh of electricity — electrolytic hydrogen plus air separation and synthesis compression (the H₂ alone is ~9 MWh/t at ~50 kWh/kg H₂ and ~0.18 t H₂/t NH₃). At ~$40/MWh (a round market anchor):

~10 MWh/t × $40/MWh ≈ $400 / t NH₃

— the dominant term in variable opex, set by the power price the system boundary exposes once electrolysis is drawn inside.

Edge case: gray ammonia consumes natural gas as both feedstock and fuel — ~30–40 GJ/t NH₃ — so its cost tracks the gas price instead, and because gas prices have swung by multiples, that one input is typically the largest sensitivity in a gray-ammonia cost.

See also

• Opex
• System boundary
• Energy intensity (per unit product)
• Fixed vs. variable cost
• Levelized cost (e.g., $/t NH₃)
• One-way sensitivity