Picture two public studies on your desk. One says a plant producing 26,000 tonnes of CSPG a year will cost about $86 million. The other puts a 44,100-tonne plant at $911 million. Capacity differs by less than 2x, yet the price tag differs by more than 10x. Which number do you trust?
The honest answer is both, and neither. Each spherical graphite feasibility study draws its process boundary differently, prices feedstock differently, and carries a different level of contingency. Line up the headline numbers and you’ll almost certainly compare apples to oranges.
This article does two things. First, it puts six public natural graphite anode studies into one table. Then it walks through, step by step, how to bring their CAPEX and OPEX onto a common basis. If you want to know what drives plant cost in general, see our guide to [What Drives the Cost of a Spherical Graphite Plant?]. Here, we stay focused on reading and comparing the studies themselves.
The Short Answer
In the five public studies that disclose plant CAPEX, a natural graphite anode material plant costs about $86 million (26,000 tpa) to $911 million (44,100 tpa). That works out to roughly 3,300–20,700 per annual tonne of capacity. Where studies report feedstock-excluded operating costs, they cluster far more tightly, at about 2,100–2,800 per tonne.
Key takeaways:
- Capital intensity, not total CAPEX, is the fair basis for comparison. It divides CAPEX by annual capacity.
- Most of the OPEX gap is feedstock. Strip it out and conversion costs differ by only about 30%.
- Contingency practice ranges from about 7% to 25%, and some studies carry risk reserves outside CAPEX entirely.
- Even after normalization, capital intensity still varies several-fold. Design maturity, location, and execution model account for much of what remains.
What These Studies Cover, and What They Leave Out
A quick note on terms. Spherical graphite (SG) is flake concentrate that has been micronized and shaped into near-spherical particles. Purified, it becomes spherical purified graphite (SPG). Coated with amorphous carbon and heat-treated, it becomes coated spherical purified graphite (CSPG), the active anode material battery makers buy. For a full comparison, see SPG vs. CSPG vs. HPG.
Most of the studies below shape the graphite first and purify it afterward. That isn’t the only workable sequence. In its natural graphite anode lines, ZDZN purifies first and shapes second, which changes how much material the purification circuit handles and where the fines end up. Keep that in mind when you map a study’s flowsheet onto your own project.
The bigger issue is what each study leaves out. Some stop at SPG and exclude coating. Some fold the mine and concentrator into the total. Some carry a risk reserve outside CAPEX. Others cover only Phase 1 or a single module. So before you read any number, pin down the boundary.
Six Public Studies Side by Side
The table covers public studies that use natural graphite as feedstock and produce spherical graphite or active anode material. We calculated capital intensity (CAPEX ÷ annual capacity) from the published figures; treat it as a rough benchmark only. All figures are in U.S. dollars as reported in each study’s base year, so they aren’t adjusted for inflation.
| Project (Location) | Published | Study Type / Class | Main Product & Annual Capacity | Initial CAPEX | Capital Intensity (approx.) | Unit OPEX & Basis |
| Falcon (Morocco) | Nov 2025 | Technical-economic study, AACE Class 3, not NI 43-101 | 26,000 t CSPG, plus fines | $86M (incl. contingency) | $3,300 per annual tonne | $2,160/t direct; $3,168/t incl. feed and transport |
| GreenRoc anode plant (Northern Europe, site TBD) | May 2024 | PFS, AACE Class 4 | 39,700 t CSPG | $321M (incl. 25% contingency) | $8,100 per annual tonne | $2,211/t (feedstock basis not stated) |
| NextSource (UAE) | Oct 2025 | Technical-economic study | 30,000 t anode material (mix of CSPG, SPG, purified flake) | $291M (incl. $13M sunk costs) | ~$9,300 per annual tonne (excl. sunk costs) | Not disclosed |
| Westwater Kellyton Phase II (U.S.) | Jan 2025 | DFS | ~37,500 t CSPG (Phase II) | $453M (incl. 20% contingency) | ~$12,100 per annual tonne | Not reported separately |
| NMG Bécancour (Canada) | Mar 2025 | Updated feasibility study, NI 43-101 | 44,100 t active anode material | $911M (materials plant only) | ~$20,700 per annual tonne | $2,810/t (mine costs reported separately) |
| Renascor (South Australia) | Aug 2023 | BAM Study (integrated mine + PSG) | Uncoated PSG | PSG portion not disclosed separately | — | $1,782/t integrated (first 10 years, incl. own concentrate cost) |
A few notes on reading each row:
- Falcon. Its estimates are built on the actual costs of a technical partner’s operating plant in China, adjusted for Moroccan logistics and construction. For background, see our coverage of Falcon’s anode plant study.
- GreenRoc. The study is a stand-alone assessment that assumes concentrate is bought at $1,100 per tonne. For the project’s latest progress, see our report on the GreenRoc pilot plant.
- NextSource. Its UAE facility will be built in two phases, starting at about 14,000 tpa for roughly $150 million. Not all of its output is CSPG.
- Westwater. The Phase II figure comes from a DFS that includes a 20% contingency. The company projects about 50,000 tonnes of total capacity across both phases.
- Renascor. Its product is uncoated PSG, and its operating cost includes concentrate valued at its own mining cost, not market price. Treat it as a separate category.
We left three projects out. Talga uses a proprietary flowsheet that differs from the conventional shaping route. NextSource’s earlier Mauritius study has been superseded by the company’s later plans. Graphite One’s Ohio plant was announced to start on synthetic graphite, which falls outside this article’s natural graphite focus.
How to Normalize the Studies: A Worked Example
The highest and lowest capital intensities in the table differ by more than 6x. Part of that gap comes from how each study is built rather than how each plant is designed. Here’s how to strip out the study effects in four steps.
Step 1: Put OPEX on the Same Feedstock Basis
Feedstock is the easiest place to go wrong. Falcon’s all-in figure of $3,168 per tonne breaks down into $2,160 of direct operating cost plus $1,008 for feed and transport. NMG reports its materials plant at $2,810 per tonne, with mine costs carried separately. GreenRoc reports $2,211 per tonne but doesn’t state whether feedstock is included. At its own assumptions of $1,100 per tonne of concentrate and about two tonnes of concentrate per tonne of CSPG, feed alone would roughly equal that figure, so it most likely excludes feedstock.
Once you compare the feedstock-excluded numbers, the projects converge:
| Project | OPEX Excluding Feedstock |
| Falcon | ~$2,160/t |
| GreenRoc | ~$2,211/t (inferred basis) |
| NMG | ~$2,810/t |
That’s a gap of about 30%, compared with more than 6x on capital intensity. In other words, conversion cost isn’t what separates these projects. CAPEX, feedstock, and yield are. For the operating side of that equation, see [内链:How to Reduce Cost per Ton in Spherical Graphite Production].
Step 2: Separate Contingency and Reserves from the Base Estimate
Contingency practices vary dramatically:
- Falcon’s $86 million includes $6 million in contingency and $5 million for escalation. That’s roughly 7% contingency.
- GreenRoc applies 25%, and Westwater’s Phase II applies 20%.
- On top of its capital estimate, NMG recommends a $122 million P50 risk reserve that is not included in its CAPEX figure.
Assume each percentage is applied on top of the base estimate and back it out. GreenRoc’s base comes to roughly $257 million, or about $6,500 per annual tonne. Westwater’s base comes to roughly $378 million, or about $10,100. Going the other direction, adding NMG’s reserve lifts its figure to about $23,400.
| Project | As Reported ($/annual t) | Adjusted ($/annual t) | Adjustment |
| Falcon | ~3,300 | ~3,300 | None (contingency already ~7%) |
| GreenRoc | ~8,100 | ~6,500 | 25% contingency removed |
| Westwater Phase II | ~12,100 | ~10,100 | 20% contingency removed |
| NMG | ~20,700 | ~23,400 | $122M risk reserve added |
These adjustments narrow some gaps and widen others, but they don’t change the order. Even on an adjusted basis, the spread between Falcon and NMG is roughly 7x. Scope, location, and design maturity, not estimating practice alone, explain the rest.
Step 3: Match the Estimate Class
Estimate class tells you how much the number can move. Falcon cites an AACE Class 3 accuracy range of –15% to +20%, which puts its $86 million anywhere from about $73 million to $103 million. GreenRoc’s PFS was prepared to AACE Class 4, which carries a wider band still. So two studies at different classes can look far apart and still overlap. To see how testwork narrows these bands, read our articles on graphite processing testwork and graphite process scale-up.
Step 4: Match Product Scope and Phase
Finally, line up what each study actually delivers. Renascor stops at uncoated PSG, so it needs a coating stage added before it can sit next to the CSPG projects. One recent research analysis does exactly this, adding estimated coating CAPEX and OPEX to studies that leave coating out, and measuring every project by capital intensity.
Phase matters too. NextSource’s first phase works out to about $10,700 per annual tonne, higher than the roughly $9,300 for its full build-out. Phasing lowers the first check you write, but not necessarily the cost per tonne of capacity.
Why Falcon’s CSPG Plant CAPEX Is So Much Lower
Falcon stands out at roughly one-sixth the capital intensity of NMG, so it’s worth looking at why. The published breakdown shows only about $30 million for the CSPG processing circuit itself, with structural steel at $22 million and the rest spread across civil works, power infrastructure, indirect costs, escalation, and contingency. Three features of the study explain much of the gap:
- It borrows a proven design. Capital and operating costs are based on a plant already running in China, adjusted for local conditions.
- It uses one main building. A single “Super Building” on about five hectares keeps the footprint and structural costs down.
- It carries thin contingency. At roughly 7%, contingency is far below GreenRoc’s 25% or Westwater’s 20%.
The study isn’t an NI 43-101 technical report, and its accuracy range runs to +20%. Still, it shows how much of a plant’s capital cost comes down to design maturity and execution model, not just the process itself.
What the Studies Reveal About Hidden Scope
Reading these studies side by side also shows where the real engineering risk sits. Three items come up again and again.
Water and off-gas treatment. Falcon specifies a water treatment system sized for about 432,000 tonnes a year, plus lime-based off-gas scrubbing. NMG names optimizing its water treatment process as its top project risk. If a study covers this scope with a single generic factor, give it a second look. For how purification routes shape these systems, see our natural graphite purification equipment selection guide.
Coating and carbonization utilities. GreenRoc notes that nitrogen bought in tanks for its coating kilns is a significant consumable and flags on-site generation as a savings opportunity. Kiln selection sets protective-gas use, temperature uniformity, and how continuous the operation can be. Rotary kiln designs offer distinct trade-offs here, covered in the critical role of the coating granulation rotary kiln.
Supplier integration. NMG lists the integration of key Asian suppliers into detailed engineering and construction as a major risk. When equipment comes from several vendors, the interfaces between them are where change orders and schedule slips start. A single party that owns the process package, equipment, and commissioning reduces that risk.
Process sequence belongs on this list too. Purify-then-shape and shape-then-purify affect purification throughput, fines handling, and overall yield differently, and the cheapest time to test that trade-off is during the study, before the design is locked.
Frequently Asked Questions
Which of these spherical graphite studies is NI 43-101 compliant? Of the studies compared here, NMG’s updated feasibility study is an NI 43-101 technical report, while Falcon’s study states explicitly that it is not one. When you weigh studies against each other, a report prepared under a recognized disclosure standard generally deserves more weight than a company-level technical study.
What is capital intensity, and why use it? Capital intensity is initial CAPEX divided by annual production capacity, expressed in dollars per annual tonne. It lets you compare plants of different sizes on one basis. Always note whether the figure covers a single phase, a module, or full build-out.
Why is Falcon’s CSPG plant so much cheaper than NMG’s? Falcon’s estimate is based on an operating plant in China, uses a single main building, and carries about 7% contingency. NMG’s is a full NI 43-101 study for a North American greenfield plant, with a separate $122 million risk reserve on top. Scope, location, and estimating practice explain much of the difference.
Does CSPG plant OPEX include graphite concentrate? Not always. Some studies list feedstock separately, some value it at the cost of self-produced concentrate, and some assume market-price purchases. Confirm the basis before you compare.
Can I use these figures to budget my own project? Only as order-of-magnitude references. Your feedstock, site, product spec, and process sequence will differ. Use the studies to understand cost structure, then build your own estimate from testwork and equipment quotes tied to specific process steps.
Conclusion
Public feasibility studies are one of the best references for spherical graphite project economics, as long as you read them on a common basis. Strip out feedstock and operating costs look surprisingly similar. Strip out contingency and reserves and capital intensity still varies several-fold, which tells you that design maturity and execution matter as much as the process itself.
Wherever your project stands, ZDZN can help:
- Still in early research? Explore our guides to planning a natural graphite processing plant and graphite concentrate qualification to build a solid foundation in process and feedstock.
- Have a study or quote that needs a second opinion? Book a free technical consultation. Our engineering team can review your process boundary, equipment configuration, and key cost assumptions using the normalization steps in this article.
- Getting ready for engineering? ZDZN provides full-line solution assessments from purification and shaping through coating, carbonization, and final product handling, along with equipment configuration, EPC delivery planning, and anode material performance testing to support process selection.