Table of Contents
In March 2021, a fire at a Renesas plant in Naka, Japan, knocked out a factory responsible for roughly two-thirds of the microcontrollers used in cars. A month earlier, a winter storm in Texas had shut fabs owned by Samsung and NXP. Within weeks, assembly lines on three continents went quiet, and “chip shortage” became a phrase people used at the kitchen table.
Five years on, the picture is messier and more interesting. Some shortages have genuinely eased. Others have simply changed address, moving from silicon to minerals, from minerals to machine tools, from machine tools to people. The automotive industry shortages of 2025 look less like a crisis with an end date and more like a permanent condition that manufacturers are learning to price in.
The chip crisis ended. The chip shortage didn’t.
Lead times for many automotive semiconductors have fallen from the 52-week peaks of 2022 back to something close to normal. That headline hides a split. Car demand is concentrated on mature nodes, chips built at roughly 28 to 90 nanometres. The enormous capacity investment of the past few years went almost entirely into leading-edge nodes for AI accelerators and phones. A 40nm microcontroller etched on a 200mm wafer gets no benefit from a $20bn 3nm fab.
The money involved was never trivial. AlixPartners put the revenue cost of the 2021 chip shortage at around $210 billion, with roughly 11 million vehicles of production lost worldwide. Suppliers still carry the scar tissue.
Why legacy nodes stay tight
Most 200mm production equipment is no longer made. Refurbished tools sell at a premium, and a new mature-node line takes two to three years to bring online with returns that make investors nervous. Meanwhile every new vehicle adds content: a mid-range model now carries somewhere between 1,000 and 1,400 chips, and the average is climbing as driver-assistance and infotainment systems multiply.
The parts nobody forecasts
Tier-2 and tier-3 suppliers remain the weak link. A $0.40 connector or a single custom ASIC can halt a €40,000 vehicle. Those small vendors rarely get the visibility that big OEMs demand, and they were the last to recover from the last crunch.
Minerals and magnets: the shortage that moved underground
Battery-grade lithium carbonate ran from around $5,000 a tonne in 2020 to above $80,000 in late 2022 before collapsing. That volatility wasn’t really a shortage of rock. It was a shortage of processing capacity, permitting and patience.
The sharper problem sits in magnets. China refines roughly 90% of the world’s rare earths into the neodymium-iron-boron magnets found inside traction motors, seat adjusters and audio speakers. When Beijing added export licensing for several magnet-related categories in April 2025, purchasing teams across Europe and India spent the summer hunting alternatives. Some lines paused over a component the size of a poker chip.
Graphite is the quiet one. China has required export permits for graphite since late 2023, and synthetic anode material for cells is tightly held. Combined with nickel and cobalt supply routed through a handful of jurisdictions, the mineral layer of the supply chain now carries more geopolitical risk than the electronics layer ever did.
Tooling, castings and the trades nobody trains for
Before a single panel is stamped, somebody has to cut a die. Large stamping dies and injection moulds routinely take 12 to 18 months from order to first part, and the number of independent tool and die shops in North America and Western Europe has shrunk for two decades. Add a new model launch and the queue gets brutal.
- Large structural castings: gigacasting capacity is concentrated in a handful of suppliers, and a single machine outage can stall a whole platform.
- High-voltage connectors and busbars: low volume, high qualification burden, few qualified plants.
- Wiring harnesses: labour-intensive, hard to automate, and vulnerable to regional disruption.
- Skilled trades: tool and die makers, maintenance technicians and HV-certified engineers sit near the top of every hard-to-fill list.
Why shortages linger even when demand cools
The bullwhip explains more than any single factory fire. When allocation was the rule in 2022, buyers doubled orders to protect themselves. That behaviour rippled down the chain, and suppliers responded by building buffers they had spent thirty years stripping out. Just-in-time became just-in-case, with automotive chip inventories rising from around 30 days of cover to 45 or 60 at many OEMs.
Buffers cost money, and those costs sit inside vehicle prices alongside tariffs, energy and labour. They are part of the compounding pressures squeezing manufacturer margins right now, and they explain why sticker prices have proved sticky even as some commodity costs fall back.
What this means on the forecourt
Consumers feel shortages through three channels: price, patience and choice. Popular trim levels disappear from order books. Features get substituted. A factory-order car that took eight weeks in 2019 can still take four months if it needs a particular powertrain or option pack.
Practical responses that actually help:
- Ask the dealer which configurations are pre-allocated to their region rather than built to order.
- Check whether a feature you want is bundled with a constrained component, such as a panoramic roof or a premium audio system.
- Compare total cost, not just list price. Higher trims often hold better residuals when supply normalises.
- For fleet buyers, lock contracts with price-adjustment clauses tied to named commodities rather than accepting open-ended increases.
There is a wider affordability story underneath all this. The shift toward software-defined vehicles adds cost at exactly the moment buyers are most price-sensitive, which is central to the current fight over what a new car should cost.
The supply chain is quietly being rebuilt
Three structural changes are underway. First, OEMs are signing direct capacity deals with chipmakers instead of buying through distributors, following agreements struck by GM, Ford and Stellantis with foundry partners. Second, cell and component plants are clustering next to assembly sites in Ontario, Kentucky, Hungary and Spain, shortening logistics and shielding factories from border shocks. Third, dual sourcing has gone from a procurement slogan to a board-level mandate, with Mexico, Vietnam and India picking up work that once went to a single low-cost country.
None of this is fast. A battery plant takes two to four years. A new magnet refinery takes longer, and the permitting alone can eat a decade.
Three things worth watching over the next 18 months
Magnet capacity outside China. Lynas, MP Materials and a cluster of European and Japanese projects are scaling separation and magnet-making. Watch the tonnage actually shipped, not the announcements.
Graphite and anode supply. Synthetic anode plants in the US and South Korea are the pinch point for cell chemistry, and qualification takes time even when the capacity exists.
Tariffs and rules of origin. Duty structures on imported components can turn a global shortage into a regional one overnight, because a part that exists somewhere is useless if it lands at the wrong border with the wrong paperwork.
For anyone buying, selling or building cars, the working assumption has changed. Scarcity is no longer an event to be survived. It is a design constraint, and the companies that treat it that way, with buffer stock, dual sourcing and honest lead times, are the ones delivering cars while their competitors explain delays.
