AI memory car prices are no longer a problem confined to PC builders and GPU addicts. Nikkei Asia reported that General Motors and BYD are already changing price expectations as the AI-driven squeeze on DRAM and other memory parts reaches the auto business.
GM now expects new-vehicle prices to rise 0.3% this year, Nikkei Asia reported, reversing its earlier view that vehicle costs would hold steady or become slightly cheaper. GM Chief Financial Officer Paul Jacobson said on the company’s earnings call that the automaker expects costs to rise by $1.5 billion to $2 billion, driven mostly by pricier vehicle components, especially DRAM.
BYD, China’s largest electric-vehicle maker, has raised prices for its driver-assistance features by 20%, according to Nikkei Asia. BYD sells those features separately from its vehicles, so the increase hits a specific software-and-hardware package rather than the base sticker price of every model.
Hyundai has also pushed for a stronger domestic chip supply chain in South Korea, Nikkei Asia reported. That is the predictable response when a car company realizes the memory market is being yanked around by AI infrastructure spending and that automotive purchasing managers do not get to wish their way into more qualified chips.
Why does an AI memory shortage affect car prices?
Modern cars use far more memory than older vehicles because infotainment, driver-assistance, safety and centralized computing systems all need DRAM and NAND to run. Automakers cannot just buy whatever memory part is cheap that week, because automotive chips need long validation cycles and must survive heat, cold, vibration and weather.
Micron said in 2023 that the average vehicle used 90GB of combined DRAM and NAND. The company expected that figure to rise to 278GB by 2026, with high-end vehicles needing as much as 2TB.
A British original equipment manufacturer cited by Nikkei Asia broke down the current appetite: infotainment systems require 4GB to 16GB of memory, advanced driver assistance and safety systems need 8GB to 32GB, and centralized systems usually start at 16GB and can reach 64GB or more. As AI assistants move into vehicles, those systems require at least 256GB of memory, according to the same report.
Micron has also forecast that Level 4 autonomous vehicles will need at least 300GB of RAM. Level 4 autonomy means a vehicle can drive itself under defined conditions without human intervention, which turns the car into a rolling compute box with tires and warranty paperwork.
The hard part for automakers is timing. Automotive memory is not interchangeable with consumer PC parts, even when the labels sound familiar. Vehicle-grade components can take months or years to qualify because failure has to be considered under road shock, temperature swings and long service lives, not just whether a desktop boots after a BIOS update.
That makes the AI memory shortage more painful for carmakers than a normal commodity price swing. If supply stays tight, buyers could face higher prices, fewer configurations or slower deliveries, echoing the availability problems the industry saw during the pandemic-era chip and logistics disruptions.
This story draws on original reporting from Tom's Hardware.
