What are the critical mineral bottlenecks (lithium, cobalt, rare earths) that could stall the energy transition
Why the Metals We Need for Clean Energy Are So Hard to Get
Based on analysis of a 126-node, 429-edge knowledge graph mapping the structural relationships between critical mineral supply chains and the global energy transition.
Imagine you are building a bicycle factory. You know you will need steel, rubber, and aluminum. You know demand is coming. But the mines that produce those materials take 16 to 20 years to open, the refining plants are mostly owned by one competitor who can set prices, and every time you try to switch to a different material, you find out that material has the same problem in a different country. That is roughly the situation the energy transition faces with lithium, cobalt, rare earths, graphite, nickel, and a handful of other minerals.
A large knowledge graph — 126 concepts, 429 connections between them — maps out how these problems relate to each other. The picture that emerges is not just “these minerals are scarce.” It is more specific: a set of interlocking structural traps, where fixing one problem tends to activate another one, and where the solutions that exist mostly arrive too late to help.
The Central Bottleneck: Mines Take a Long Time
The single most connected node in the entire graph is called the Mining Lead Time Trap. It shows up in 58 different relationships and sits at the center of the whole structure.
Here is what it means. If you want to open a new mine today, the typical timeline from discovery to production is 16 to 20 years. That is not a policy choice or a funding shortfall — it is the physical reality of drilling, permitting, building infrastructure, and commissioning equipment. It cannot easily be compressed.
This matters enormously because everything else in the graph feeds into this bottleneck. When prices for lithium or cobalt drop (discouraging investment), that amplifies the lead time trap. When Western countries take years to issue mining permits, that amplifies it. When AI data centers add unexpected demand for grid-scale batteries, that amplifies it. When a price crash causes investors to pull funding from junior mining companies, that amplifies it. Each of these is an independent pressure, but they all funnel into the same physical constraint: mines take a long time to build no matter how urgent the need is.
The graph identifies this node as both an outcome (many things make it worse) and a cause (it then constrains supply, drives price cycles, and blocks policy goals). It is the point at which all the different pressures become one shared physical problem.
Switching Chemicals Does Not Escape the Trap
One obvious response to mineral concentration is to change the recipe. If cobalt from the Democratic Republic of Congo is the problem, switch to lithium iron phosphate (LFP) batteries, which contain no cobalt. The graph tracks what actually happens when you do this.
LFP batteries do reduce dependence on DRC cobalt. That part works. But LFP batteries require phosphate — and Morocco holds 60 to 70 percent of global phosphate reserves. The graph records a direct edge: LFP Chemistry Cobalt Bypass creates new dependency at Morocco Phosphate LFP Chokepoint, with one of the highest relationship weights in the entire graph. You have moved the concentration from one country to another.
LFP batteries also still require graphite for their anodes. China controls over 90 percent of battery-grade graphite processing. The graph records this explicitly: LFP Chemistry Cobalt Bypass fails to escape Graphite Anode China Monopoly.
The same pattern repeats when you look at the next generation of chemistries. Manganese-based batteries (LMFP) are being developed partly to reduce dependence on the materials above. The graph has a node called Manganese Battery Grade Processing Chokepoint — the processing of high-purity manganese is already concentrating in similar ways to prior chokepoints.
This is what the graph calls the Mineral Substitution Cascade Effect. Each chemistry transition displaces one concentration while activating another. The chair moves; the game of musical chairs does not end.
The Layer Nobody Talks About: Refining
Even if you diversify mining — opening lithium mines in Australia, nickel mines in Indonesia, cobalt mines in other African countries — there is a second layer of concentration that mining diversification does not address.
Between a raw ore and a battery cell, there is a refining and processing step. Lithium ore needs to be converted into lithium hydroxide. Graphite needs to be processed to battery-grade purity. Nickel ore from Indonesia needs to go through a technically difficult process called HPAL (high-pressure acid leach) to become battery-grade nickel sulfate.
China controls this processing layer comprehensively. The graph treats this as a structurally separate problem from mine ownership. A node called China Battery Materials Midstream Monopoly has both high connectivity and high weight, meaning the graph identifies it as independently important — not just a consequence of something else.
The practical implication: you can own a lithium mine outside China and still depend entirely on Chinese processing capacity to turn your ore into something a battery manufacturer can use. The graph records this explicitly in the case of Indonesia, where export nationalism rules were designed to force domestic value-added processing — but Chinese firms ended up controlling the functional HPAL capacity that was built.
Why Policies Keep Getting Blocked
The United States passed the Inflation Reduction Act (IRA), which includes provisions designed to reduce dependence on Chinese mineral processing by restricting tax credits for EVs that use materials from Chinese-controlled supply chains. The graph records what happens next.
The IRA restrictions on Chinese-sourced materials — specifically the FEOC (Foreign Entity of Concern) provisions — are blocked by the same Mining Lead Time Trap they are trying to address. If American manufacturers cannot qualify for EV tax credits unless they source graphite from non-Chinese processors, but no non-Chinese graphite processing capacity exists yet, and building it takes 16 to 20 years, the policy produces a gap in qualified EVs rather than a shift in supply chains.
The graph records three separate pathways by which the IRA’s primary instrument is structurally blocked: by physical lead times, by a 2025 policy reversal that reduced funding and certainty, and by what the graph calls the IRA Friend-Shoring Effectiveness Gap — the network of allied countries with potential supply is not yet producing at the volumes needed to substitute for Chinese processing.
This is not an argument about whether the policy is good or bad. It is a structural observation: the policy’s timeline and the physical reality of mine and plant development are mismatched, and the graph contains no resolution mechanism that closes this gap.
When Problems Reinforce Themselves
The graph identifies several feedback loops — situations where a problem makes itself worse over time.
The most important one involves investment. When lithium prices drop (as they did sharply in 2022-2024), mining companies reduce investment in new projects. Fewer new projects means less future supply. Less future supply means prices will eventually spike. The spike triggers new investment — but those projects take 16 to 20 years to produce. By the time they come online, the market may have shifted again.
The graph records evidence that China’s industrial policy deliberately exploits this cycle: Chinese state-owned enterprises expanded lithium production aggressively during the 2021-2022 price spike, driving prices down and causing Western junior mining companies (smaller exploration firms that fund early-stage projects) to lose financing. This is the mechanism called China Mineral Price Suppression Weapon in the graph. The long-run effect is to reduce Western investment capacity while Chinese companies — which do not face the same commercial return requirements — continue expanding.
Some Non-Obvious Things the Graph Shows
A few findings in the graph are genuinely counterintuitive.
Sodium-ion batteries have been proposed as a way to escape mineral concentration — they do not require lithium, cobalt, or graphite. The graph confirms this: sodium-ion batteries do bypass those specific mineral chokepoints. But they deepen Chinese manufacturing concentration, because CATL (a Chinese company) is the dominant developer and manufacturer of sodium-ion cells. The mineral liberation is real. The manufacturing liberation is not.
Indonesia banned raw nickel exports in 2020 to force battery-grade processing to happen domestically. The graph traces what happened: the export ban inspired the DRC to try a similar tactic with cobalt. Cobalt weaponization accelerated the shift to LFP batteries. LFP batteries reduced nickel demand. Indonesian HPAL plants — the processing facilities the export ban was supposed to create — ended up controlled by Chinese firms. The stated goal of capturing downstream value was achieved by Chinese processors rather than Indonesia.
Recycling is often described as a long-term solution. The graph shows it is — but with a structural timing problem. Battery recycling provides meaningful material recovery, but only from batteries that have already reached end of life. The first large wave of electric vehicles was sold in the early 2020s, with typical battery lifespans of 10-15 years. That puts substantial recycling supply online in the mid-to-late 2030s. The graph records this explicitly: Battery Recycling Circular Supply partially resolves after 2035 Mining Lead Time Trap. Recycling is real, but it is structurally too late to address the supply gap in the near term.
AI data centers are not typically part of conversations about battery minerals, but the graph records a direct amplification pathway. Large data centers require significant battery backup systems, and the rapid growth of AI infrastructure adds a demand vector that was not included in EV-only projections.
Bottom Line
The graph’s structural picture can be summarized in four observations.
First, the Mining Lead Time Trap is not one problem among many — it is the point at which all other pressures become a shared physical constraint. Regulatory barriers, price crashes, investment gaps, environmental conflicts, and demand surges all funnel into the same bottleneck: mines take a long time to build.
Second, chemistry substitution is real but incomplete. Switching battery chemistries reduces dependence on specific minerals, but each transition tends to activate a new concentration in a different material or country. The overall level of supply chain concentration does not obviously decrease.
Third, midstream processing is a structurally distinct layer from mining. Supply chain diversification that reaches only as far as the mine does not address the refining and processing concentration that sits between ore and battery cell.
Fourth, the most prominent policy response (the IRA’s FEOC provisions) and the most prominent alternative supply mechanisms (recycling, seabed mining, chemistry substitution) all face structural timing mismatches with the near-term supply gaps the graph identifies. The graph does not contain resolution mechanisms that close these gaps before the mid-2030s.
The graph is not a prediction. It is a map of structural relationships as they existed when the data was collected. What it shows is that the obstacles to critical mineral supply are not independent problems — they are an interconnected system, and addressing any one of them without accounting for the others tends to move the constraint rather than remove it.