The standard narrative of the EV transition runs as follows: we are escaping oil dependency, reducing geopolitical exposure, cleaning the energy system. Each of these claims contains truth and each obscures something larger. The transition substitutes one legible dependency for roughly twenty less legible ones, adds a control layer embedded in the vehicle itself that petroleum never possessed, and hands three distinct sets of powerful actors — not one — significant new leverage over how people move. The map being sold is missing most of the terrain.
I. The Upstream Tree
The petrol car’s upstream dependency was shallow: crude oil, steel, aluminium, rubber, a thin electronics layer. Three or four geopolitical chokepoints, all named, all mapped. The electric vehicle has three parallel dependency stacks where the ICE vehicle had one — battery, motor, and infrastructure — each branching into its own chain of materials, processing stages, and geopolitical exposures.
The battery chain alone runs seven levels deep:
DIAGRAM A — The Battery Chain
ELECTRIC VEHICLE
│
▼
BATTERY PACK
(assembly — Europe manages this)
│
▼
BATTERY CELLS ◄─── Northvolt failed here
(electrochemical manufacturing)
│
▼
ACTIVE MATERIALS
(cathode/anode powders)
~80% China
│
▼
REFINED CHEMICALS
lithium hydroxide · cobalt sulphate · spherical graphite
65–95% China
│
▼
RAW MATERIALS
Lithium ── Chile, Argentina, Bolivia
Cobalt ─── DR Congo (70% of world supply)
Graphite ── global mining, but 95% processed in China
Nickel ──── Indonesia, Philippines, Russia
│
▼
REAGENTS
sulphuric acid (fossil fuel byproduct)
sodium cyanide · organophosphorus extractants
───────────────────────────────────────────────
KEY: Each arrow down = one processing stage
China controls the middle four layers
Europe controls only the top
───────────────────────────────────────────────
The electric motor adds a separate chain entirely — permanent magnets requiring rare earth elements (neodymium, dysprosium) of which China controls roughly 90% of both mining and processing. This is a distinct chokepoint from the battery chain, rarely discussed alongside it.1
II. The Chokepoints
The vehicle also depends on infrastructure that did not exist as a dependency before — a charging network, a grid, and the electricity generation beneath it:
DIAGRAM B — The Infrastructure Chain
ELECTRIC VEHICLE
│
▼
CHARGING NETWORK
│
▼
GRID (local distribution)
╱ ╲
▼ ▼
EQUIPMENT PERMITTING
Transformers Avg. 5+ years
18–36 month delivery to connection
Inverters, grid-scale ~80% of projects
storage: significant never reach
Chinese manufacture construction
╲ ╱
▼ ▼
ELECTRICITY GENERATION
╱ │ ╲
▼ ▼ ▼
RENEWABLES NUCLEAR GAS
Inverters: New builds: LNG market
Chinese supply decade-scale fills gap
Wind magnets: timelines regardless
REE (90% SMRs: no of policy
China) commercial
deployment
anywhere yet
───────────────────────────────────────────────
KEY: Left bottleneck = geopolitical (Chinese equipment)
Right bottleneck = self-inflicted (European permitting)
Generation layer = no clean solution in medium term
───────────────────────────────────────────────
The two bottlenecks in the grid node are different kinds of problem. The equipment chokepoint is geopolitical — transformers, inverters, and grid-scale storage manufactured predominantly in China, meaning the infrastructure being built to reduce energy dependency is itself dependency-laden. The permitting chokepoint is self-inflicted — European planning systems, administrative capacity, and project approval processes that fail 80% of proposed projects before construction begins. One requires industrial policy to fix. The other requires political will to fix. Neither is being addressed at the required speed.
Beneath the grid sits the generation problem. Renewables cannot provide the firm baseload that an electrified economy — powering both EV fleets and AI data centres simultaneously — requires. New nuclear is necessary but slow; small modular reactors have no commercial deployment anywhere and realistic European timelines put first operations in the mid-2030s. Gas fills the gap, whether stated openly or not. The dependency has not been eliminated. It has been rerouted.
Seen together, the three chains produce a dependency map that the petrol car never had:
DIAGRAM C — The Full Dependency Tree
ELECTRIC VEHICLE
│
├─── BATTERY PACK (assembly — Europe manages this)
│ │
│ └─── BATTERY CELLS (manufacturing — Northvolt failed here)
│ │
│ └─── ACTIVE MATERIALS (cathode/anode powders — ~80% China)
│ │
│ └─── REFINED CHEMICALS
│ (lithium hydroxide, cobalt sulphate,
│ spherical graphite — 65–95% China)
│ │
│ └─── RAW MATERIALS
│ Lithium: Chile, Argentina, Bolivia
│ Cobalt: DR Congo (70%)
│ Graphite: global, but processed
│ in China (95%)
│ │
│ └─── REAGENTS ¹
│ (sulphuric acid —
│ fossil fuel byproduct)
│
├─── ELECTRIC MOTOR
│ │
│ └─── PERMANENT MAGNETS
│ │
│ └─── RARE EARTH ELEMENTS
│ (neodymium, dysprosium — ~90% China,
│ mining AND processing)
│
└─── INFRASTRUCTURE
│
├─── CHARGING NETWORK
│ │
│ └─── GRID (distribution)
│ │
│ ├─── EQUIPMENT
│ │ (transformers — 18–36 month delivery,
│ │ inverters, grid-scale storage —
│ │ significant Chinese manufacture)
│ │
│ └─── PERMITTING
│ (avg. 5+ years to connection;
│ ~80% of projects never built)
│
└─── ELECTRICITY GENERATION
│
├─── RENEWABLES (inverters — Chinese supply;
│ wind turbine magnets = REE)
├─── NUCLEAR (uranium — Kazakhstan, Russia, Canada;
│ new builds: decade-scale timelines)
└─── GAS (LNG — global market dependency;
fills gap regardless of policy)
SOFTWARE & DATA LAYER (sits across entire tree)
─────────────────────────────────────────────────
Journey logging · Charging authentication
OTA updates · Platform control (Tesla, Google, Apple)
Data: commercial value · regulatory value · control capability
III. The Third Player
The standard political framing pits fossil fuel incumbents against electrification advocates. This misses the actor who gains most from the transition regardless of who wins the energy competition.
Platform economy companies — Tesla, Google, Apple, Microsoft — are not primarily interested in battery chemistry or grid topology. They are interested in the software and data layer sitting across the top of the entire dependency tree. Every EV is a connected device that logs journeys, authenticates charging sessions, and receives over-the-air updates that can modify vehicle behaviour after purchase — adding features, removing them, adjusting range, changing charging parameters. This has happened, on purchased vehicles, without owner consent, in documented cases.2 The legal framework governing what manufacturers and platform operators can do with this capability is almost entirely unresolved in European law.
The structural parallel is China’s cashless payment infrastructure — not as conspiracy theory but as proof of concept. WeChat Pay and Alipay were built for convenience and adopted voluntarily. The consequence is that economic participation can be suspended for an individual with a single administrative decision. The EV ecosystem is the mobility equivalent: infrastructure built for convenience that creates, as a structural byproduct, the capability for control. Whether that control is exercised by platform companies monetising journey data, grid operators managing demand response, or governments with other intentions is an open question. That the capability is being embedded in a continent’s vehicle fleet before governance exists to constrain it is not.
Three players are shaping the transition simultaneously: legacy hydrocarbon incumbents managing decline, industrial policy actors building national supply chains, and platform economy companies capturing the connected layer on top. The consumer and the climate sit downstream of all three — occasionally served, occasionally not, depending on which interest dominates in a given policy moment.
Closing
The EV transition trades one upstream dependency for roughly twenty, most of them more geographically concentrated than crude oil ever was and less legible to the public that bears the exposure. It adds a permitting bottleneck that is entirely self-inflicted and a control layer embedded in the vehicle itself that the fossil fuel era never possessed. The direction may still be right — the carbon arithmetic, over a long enough horizon, favours electrification. But a transition entered with open eyes about what is being exchanged can be governed intelligently. The romantic version cannot.
This essay extends arguments developed in “Clean on Paper: The Bill Behind the EV Transition” (June 2026).
Disclaimer:
Written in collaboration with Gemini 3.5 Flash
A further dependency layer beneath the materials themselves — the chemical reagents required to process them — is almost entirely absent from public discussion. Sulphuric acid is the master input for hydrometallurgical processing of copper, nickel, cobalt, lithium, and rare earth elements. More than half of global sulphuric acid production is consumed by the fertiliser industry, creating a direct molecule-for-molecule competition between food production and metal processing. The acid is predominantly a byproduct of fossil fuel refining — meaning the transition’s chemical substrate is structurally dependent on what the transition is trying to replace. Two converging shocks in early 2026 — the effective closure of the Strait of Hormuz to sulphur shipments and China’s comprehensive sulphuric acid export ban — turned this latent vulnerability into an immediate supply crisis. See Craig Tindale, The Global Reagent Squeeze (May 2026).
Tesla has modified purchased vehicles remotely on multiple documented occasions: reducing battery range on older Model S vehicles via software update without owner consent, adjusting Supercharger speeds, and enabling or disabling paid features after purchase. In 2023, a Belgian court ruled against Tesla in a case involving unilateral range reduction. The broader legal question — what manufacturers are permitted to do to vehicles after sale via over-the-air updates — remains unresolved across most European jurisdictions.
Thank you for this article. We all need to press for more governance on this and similar issues before it runs away from us and we lose the opportunity to have any say at all.