Insight, Minerals & Industrial Policy
The DRC is building a US$2 billion zone to make battery precursor materials rather than export cobalt. The barrier is not power or skills. It is qualification, a process that takes years and locks incumbents in. The same mechanism that keeps the door shut is the reason it is worth walking through, and it may not stay open.
The Democratic Republic of Congo is advancing the Musompo Special Economic Zone in Lualaba (as reported), a roughly 900 hectare site targeting around US$2 billion of private investment, designed around domestic production of nickel manganese cobalt battery precursor materials rather than the export of cobalt.
We wrote earlier this week about the move from concentrate to refined metal, and about firm power being the constraint that has now found an answer. Precursor manufacturing is a different proposition again, and the constraint on it is one that almost never appears in African industrial policy discussion.
It is not electricity. It is not skills, chemical inputs, logistics or customers, although all of those are real and all appear in the standard list. It is qualification, and understanding what qualification does explains both why this is so hard and why it is worth attempting.
The mental model of a value chain as a ladder, with each rung slightly higher than the last, is actively misleading here.
Mining and concentration are physical processes. Crushing, grinding, flotation, separation. The product is judged on grade and impurity limits, and the business is a volume business run by mining engineers.
Precursor cathode active material is a specialty chemicals business. Producing nickel manganese cobalt hydroxide involves controlled co-precipitation in stirred reactors, where particle size distribution, morphology, tap density, sphericity, internal porosity and trace element content all have to be held within narrow tolerances, batch after batch, because each of them affects how the finished cell performs and degrades over years.
The gap between those two businesses is not a rung. It is a change of industry, of workforce, of quality management culture and of customer relationship. A country that is excellent at mining has demonstrated very little about its ability to do this, and the same was true of every country that has made the transition successfully.
Here is the mechanism that governs entry, and it is worth setting out in full because it does not appear in most industrial strategy documents.
A cell manufacturer cannot simply buy precursor on specification. Cathode performance is sensitive to material characteristics in ways that are not fully captured by any written specification, so a new supplier has to be qualified. That process runs roughly as follows: sample quantities, then analytical characterisation, then small cell builds, then electrochemical cycling over hundreds or thousands of cycles to establish degradation behaviour, then pilot line trials, then a scale-up validation, then a limited commercial supply agreement with continued monitoring. Any change in the supplier's process, feedstock or plant afterwards can require partial requalification.
This takes years, not months. It costs both parties real money. And once a supplier is qualified, the cell manufacturer has a strong reason not to switch, because switching means paying the cost again and accepting the risk that a proven cell design behaves differently.
That is the definition of a lock-in, and its effects run in both directions. Qualified incumbents are locked in as suppliers. New entrants are locked out, not by tariffs or capability but by the calendar and by the risk aversion of a customer whose warranty exposure runs to a decade.
W. Brian Arthur's 1989 account of competing technologies and lock-in identified four sources of increasing returns that cause an early lead to become self-reinforcing. Precursor supply exhibits all four, which is unusual and worth walking through, because it explains why the current concentration of the industry is stable rather than accidental.
Large set-up costs. A precursor plant is capital intensive, and the fixed cost is spread over volume, so an incumbent with scale has a unit cost the entrant cannot match at start-up.
Learning effects. Yield, consistency and impurity control improve with cumulative production. An incumbent that has run millions of tonnes has a quality record a new plant cannot demonstrate, and the quality record is what qualification tests.
Coordination effects. Cell designs, equipment, testing protocols and engineering conventions have co-evolved around the materials that were available. New entrants must fit an ecosystem shaped by incumbents.
Adaptive expectations. A purchaser choosing a supplier is partly forecasting who will still be supplying in ten years. Uncertainty about a new entrant is itself a reason to prefer the incumbent, which makes the expectation self-fulfilling.
Paul David's account of QWERTY makes the general point that the outcome need not be efficient. A configuration can persist because of the sequence in which it was adopted rather than because it is the best available, and it can persist long after the conditions that produced it have changed.
Applied here: the concentration of precursor and cathode active material production in China is not a natural consequence of geology, and it is not primarily a cost phenomenon. It is a path dependent outcome of deliberate investment made when the industry was young and qualification relationships had not yet formed. Chemistry World's account of the lithium iron phosphate story makes the same point about a different chemistry: strategic government support, early licensing access and large scale manufacturing allowed China to commercialise LFP rapidly while Western countries missed early opportunities.
Somebody entered when the window was open. That is the whole mechanism, and it is the reason Musompo is worth taking seriously rather than dismissing.
If qualification is the binding constraint, the standard special economic zone playbook is aimed at the wrong target. Land, power, tax holidays and one stop permitting reduce the cost of building the plant. They do nothing about the years between commissioning and the first commercial order, which is where a precursor entrant dies.
Three implications follow, and they are specific.
A cell manufacturer has to be a partner at design stage, not a customer found later. Qualification is a joint process. It requires the buyer's laboratories, cycling capacity and engineering attention. A zone that builds first and seeks offtake afterwards has committed the capital before establishing whether anyone will do the work of qualifying it. The single most valuable thing the DRC could secure at Musompo is not two billion dollars of investment. It is one cell manufacturer contractually committed to a co-qualification programme from the start.
The financing has to survive the gap. Conventional project finance expects revenue to begin at commissioning. Here, commissioning is the start of a multi-year unpaid process. That is a concessional finance problem of a very specific shape, and it is not the shape blended finance instruments are usually built to address.
Consistency, not capacity, is the deliverable. The plant's value is demonstrated by producing identical material for two years, not by producing a great deal of it. That has implications for how the plant is designed, staffed and measured, and it sits uncomfortably with the way industrial zones report progress.
Path dependence has a corollary that is easy to state and hard to act on. If lock-in follows entry, then the value of entering depends entirely on whether the configuration you are entering will still matter.
Two developments should worry anyone underwriting an NMC precursor plant.
The chemistry is losing share. The IEA reports that lithium iron phosphate batteries now supply almost half the global electric car market, up from less than 10 per cent in 2020, at the expense of the previously dominant nickel based NMC chemistries, driven by improved performance and lower costs. LFP also accounted for over 90 per cent of global stationary battery storage installations in 2025. NMC retains a substantial position, particularly at the premium and long range end, where high nickel variants together with NCA accounted for roughly 80 per cent of 2025 electric vehicle battery deployment using cobalt containing chemistries. But the direction is not ambiguous, and LFP contains neither cobalt nor nickel.
The step itself may be disappearing. This is the more serious risk and it is barely discussed. The IEA notes that processes producing cathode active material without using precursors are being developed, which could reduce exposure to geographically concentrated supply chains, though they pose greater challenges in ensuring the required consistency in material characteristics. A precursor plant is an investment in a stage of the process that some producers are working to eliminate.
So the DRC is contemplating entering an industry in order to escape one lock-in, and the entry itself would create a new dependence on a chemistry losing share and on a process step that may not survive.
That is not an argument against Musompo. It is an argument for how the decision should be framed. The question is not whether the DRC can make precursor material. It is whether the window in which qualification relationships are still being formed for non-Chinese supply, opened by industrial policy in the United States and Europe rather than by market forces, will stay open long enough for a plant commissioned in the early 2030s to get through qualification and earn a return.
That is a judgement about policy durability and about chemistry timelines. It is answerable, approximately, and it should be answered explicitly rather than assumed.
Lualaba is not an empty industrial site. It is a landscape of industrial and artisanal mining with a long and contested history around land, water contamination, tailings, labour conditions and the presence of children in supply chains that customers are legally obliged to scrutinise.
A precursor plant introduces chemistry to that setting: sulphuric acid, ammonia, caustic soda, sodium sulphate effluent at volume, and a wastewater problem meaningfully harder than a concentrator's. The environmental and social terms of that are being set now, in the design of a zone, and they will be examined closely by exactly the European and North American customers whose qualification the project depends on. Under the EU battery regulation and comparable due diligence regimes, the social conditions of production are not a reputational matter but a market access one.
Which produces a convenient alignment worth naming. Doing the social and environmental work properly at Musompo is not a cost imposed on the commercial case. It is part of the commercial case, because the customer whose qualification programme the entire investment depends on will be required to audit it. The terms an industrial operator brings with it into a region are, in this instance, also the terms of the sale.
The Lab works on this in local manufacturing and climate and ecosystems, across the regions where processing capacity is being built for the first time.
If you are assessing an industrial zone whose viability depends on a customer relationship that does not exist yet, tell us what you need to know.
This is an independent insight piece by Transitions Lab. For the Lab's applied work, see Local Manufacturing & Supply Chains and Climate & Ecosystems. See also The Value Is in the Second Step on firm power and the move up the value chain, What an Incumbent Brings With It on the terms an industrial operator carries into a new region, The Ban Is Not the Policy on the export restriction that only bites once the plant and its access terms exist, A Recycled Material Is a Promise on the same qualification barrier in secondary materials, and Who Pays Decides What Gets Built on path dependence running on a demand signal rather than a policy one. To discuss a study, see Contact.