Table of Contents
Building cars used to be a fairly predictable business. You designed a platform, amortised the tooling over seven years, squeezed suppliers each spring, and shipped metal. That rhythm is gone. Between 2020 and 2025 the industry absorbed semiconductor shortages, a 60% swing in battery material prices, and a regulatory timetable that shifted twice. The automotive industry challenges of the mid-2020s aren’t one crisis. They’re a stack of them landing on the same balance sheets at the same time.
What makes this period unusual is the compounding. An electric car that’s hard to build profitably also has weaker residuals, which pushes up monthly lease prices, which cools demand, which leaves factories running at 60% utilisation. Each problem feeds the next.
The EV transition is running on someone else’s budget
Most legacy manufacturers signed off their electrification plans on three assumptions: packs would hit $100 per kWh by 2024, buyers would follow the incentives, and software would be something you bolt on. None of the three held up.
Battery costs stopped falling
Lithium carbonate traded near 60,000 yuan per tonne in early 2021, broke 590,000 by late 2022, then fell back under 100,000 during 2024. That whipsaw destroyed multi-year supply contracts and turned procurement teams into commodity traders. Cell-to-pack designs and the rise of LFP chemistry (roughly 40% of global packs now) have helped contain cost per kilowatt-hour, but the raw material volatility hasn’t gone away. Car makers are now hedging lithium the way airlines hedge jet fuel.
Residual values became a board-level problem
Three-year-old EVs lost around 30% of their value in a single year across the UK and US markets in 2024. Leasing companies repriced, and monthly payments on some models climbed even as list prices were cut. Hertz offloaded 20,000 EVs from its US fleet, citing collision repair costs and depreciation. When the people who own the cars after you stop believing in the resale curve, the whole financing model wobbles.
Software is now the hardest engineering problem in the car
A modern vehicle runs well over 100 million lines of code. That’s more than a fighter jet, and it’s written by teams that historically reported to mechanical engineering departments.
Compute has to survive a dashboard in August
Running capable AI inside a vehicle is a genuinely awkward hardware problem: limited power budget, no reliable cloud connection, and a thermal envelope that punishes anything ambitious. The constraints described in what it takes to run an LLM on a device apply almost word for word to the cabin of a family SUV. Chip designers are chasing the same low-power inference silicon, in the same spirit as NASA’s new AI space chip aimed at letting spacecraft make decisions without a ground link.
The gap between the demo and the driveway
Volkswagen’s Cariad unit burned through billions and shipped late. GM pulled Apple CarPlay and Android Auto from its EVs to own the interface, then faced customer backlash. Several brands have paused features that were promised at launch and never delivered over the air. Recalls now arrive as software patches, which is cheaper than dealership visits but harder to verify and harder to regulate.
Supply chains never reset to default
Just-in-time survived the pandemic on paper but not in practice. Tier-one suppliers still hold more inventory, dual-source more components, and quote longer lead times than they did in 2019. Smaller suppliers absorbed the cost of that buffer and many didn’t survive; European component makers have shed tens of thousands of jobs since 2023.
Managing that complexity needs systems that actually reflect how material moves. Manufacturers rebuilding their planning stack are leaning on frameworks like this practical playbook for choosing and implementing ERP, because spreadsheets stop working the moment you have three plants and a dozen country-specific compliance rules.
Tariffs redrew the map in twelve months
The US imposed a 100% duty on Chinese-built EVs in September 2024. The EU followed with countervailing duties reaching 45.3%. Canada matched the US at 100%, and Turkey, Brazil and India tightened import rules in parallel. Chinese manufacturers responded by buying land rather than shipping containers: BYD in Hungary and Thailand, Chery in Spain, several groups circling Mexico.
For everyone else, the maths got uglier. A platform designed for one market now needs regional variants, regional battery supply, and regional content rules to qualify for incentives. That’s capital nobody planned to spend twice.
The affordability wall is the real bottleneck
Demand hasn’t collapsed because people stopped wanting cars. It stalled because the total cost of ownership climbed faster than wages. The pressures stack up quickly:
- Insurance: premiums in the UK rose more than 25% in a single year, driven partly by the cost of replacing sensors embedded in bumpers.
- Repair complexity: a minor front-end collision on a vehicle with radar and lidar can run four figures before paint.
- Finance costs: higher interest rates added real money to PCP and lease payments even on unchanged list prices.
- Regulatory penalties: the UK’s ZEV mandate fines manufacturers £15,000 per non-compliant vehicle, and that cost finds its way into the sticker.
- Depreciation risk: buyers unsure about battery health and future tech simply wait.
The distribution model is cracking too
Franchised dealers spent a century as the industry’s shock absorber. Now manufacturers want fixed prices, direct orders and control of the customer relationship, while dealers carry the inventory risk and the training costs for technology they didn’t design. Agency models in Europe and direct sales in the US have produced lawsuits, buyouts and a lot of unhappy middlemen.
The transition also hits the workshop. Technicians trained on cambelts now need high-voltage certification, and the independent garage sector is facing an investment bill it can’t easily fund.
What the stronger players are doing differently
The manufacturers holding up best share a few habits. They’ve stopped treating every market as identical and cut model complexity, sometimes brutally. They’ve signed direct battery supply deals and taken equity in mines, accepting lower returns for predictability. They’ve moved software into a single accountable team with a shipping deadline, rather than a committee. And they’ve been willing to lose money on EVs for a defined period instead of pretending the margin was there.
None of that fixes the structural tension: capital spending on electrification is enormous, while the profits funding it still come largely from combustion vehicles that regulators are steadily taxing out of existence. That mismatch is the defining financial story of the sector, and it’s why analysts now value car makers on software revenue and battery contracts rather than unit volume. Anyone building exposure to the theme will find the same logic echoed in this look at the top 7 stocks to invest in 2026, where suppliers and chipmakers sit closer to the centre of the story than the badge on the bonnet.
The companies that come through this decade intact will be the ones that treated the transition as an operations problem rather than a marketing one. Cheaper batteries help. A working software stack helps more. But the deciding factor is usually whether the people running the plant floor and the people signing the supply contracts are looking at the same numbers.
