10 October 2025
A new semiconductor tug of war: AI data centers vs. automakers
By Hrishikesh S
Article Summary
AI data centers now dominate the demand for semiconductor chips. Understand what this means for automakers and the impact on the semiconductor supply chain.
The balance of semiconductor demand is shifting. For much of the past decade, the automotive industry was considered one of the strongest growth drivers for chipmakers, driven by electrification and digitalization. However, that momentum has slowed as EV adoption cools and several software-defined vehicle (SDV) programs are postponed.
Meanwhile, the boom in AI and cloud data centers is generating relentless demand for high-performance chips, and stakeholders are paying higher prices to secure them.
This divergence is reshaping how semiconductor capacity is allocated. AI data center customers now command priority, leaving other industries — including automotive — to compete for a smaller share of available supply.
AI data centers’ demand reorders the semiconductor supply chain
As electric vehicle adoption slows, overall semiconductor demand from automakers has softened, since EVs use more chips than internal combustion engine vehicles. Tariffs and regional trade frictions are further weighing on global production volumes, with S&P Global Mobility estimating that up to 600,000 fewer vehicles may be built in 2026.
In response, several manufacturers, particularly in North America, are deferring or scaling back next-generation SDV programs that would have driven higher semiconductor content.
At the same time, the chip industry's focus is shifting toward AI and cloud data centers. These hyperscale operators—companies such as Google, Meta, and Amazon—are expanding rapidly and purchasing large quantities of advanced semiconductors at premium prices. As a result, chip suppliers are prioritizing data center demand, redirecting wafer and packaging capacity toward higher-margin products.
For the automotive sector, this does not translate into an immediate chip shortage but rather into tighter allocation, longer lead times, and higher pricing pressure—particularly in memory and power components. Automakers now face the challenge of managing supply in a market where total capacity may be growing, yet their share is shrinking.
Automotive memory supply tightens as AI data centers take priority
AI data centers are also reshaping the memory market, diverting capacity from other industries. Hyperscalers and cloud providers are buying large volumes of advanced memory, pushing semiconductor production and packaging capacity toward high-bandwidth and server-grade products.
Major suppliers like SK hynix, Samsung and Micron are prioritizing higher-margin AI and server orders, reducing availability for automotive-grade dynamic random-access memory (DRAM) and NAND. Consequently, OEMs much compete based on price, lead time and allocation.
All three manufacturers also maintain dedicated automotive product lines but have less incentive to fulfill these orders. Suppliers are pausing new quotes or selectively hiking prices amid elevated AI demand.
Micron, for example, has signaled broad DRAM price increases across markets, with automotive and industrial-grade products facing the sharpest adjustments. Company executives say these moves reflect rising production and service costs for automotive memory, where returns have lagged.
Market sources suggest that while general DRAM prices may rise by 20–30%, automotive-qualified DRAM could see increases of up to 70%. This shift underscores the pricing leverage that memory vendors now hold amid tightening supply and strong demand.
As a result of DRAM price increases, automotive buyers using older nodes or automotive-qualified memory are encountering deep volatility. OEMs may need to pay higher prices, accept longer wait times or redesign systems to use alternative memory solutions.
Leverage our market intelligence and enhanced forecasting tools to make informed decisions in the evolving autonomous vehicle landscape.
Power discrete components and packaging bottlenecks squeeze automakers
AI and cloud data centers are driving massive growth in global power demand, with power consumption projected to rise from 280 TWh in 2024 to as much as 530 TWh by 2028.
Meeting this demand requires more advanced and efficient power electronics. Power discretes — standalone semiconductor components such as metal-oxide-semiconductor field-effect transistors (MOSFETs) and insulated gate bipolar transistors (IGBTs) — are increasingly used in AI data center power supply units, uninterrupted power supplies and cooling systems. Their ability to operate at high frequencies, minimize conduction losses and maintain superior thermal performance is critical for handling growing workloads while controlling energy consumption.
Data center operators are buying large volumes of advanced power components, and suppliers are prioritizing them over other customers. This leaves fewer components for the automotive sector, as the same semiconductor lines and packaging resources are shared across applications.
The key bottleneck is increasingly in back-end packaging and assembly rather than wafer fabrication. Suppliers’ focus on AI data center applications leads to slower delivery and reduced capacity allocations for automotive-grade power MOSFETs, IGBTs and power management ICs (PMICs). Despite an increase in wafer fabrication capacity, the limited packaging capacity continues to place supply pressures on automakers.
Demand for advanced back-end packaging solutions is outpacing available capacity and investment in packaging equipment. Critical packaging components are also in short supply. Bringing new packaging vendors online can take more than 25 weeks and requires significant investment, causing persistent delays in power discrete shipments.
Supply chain implications and outlook
OEMs are increasingly aware that semiconductor shortages could emerge by late 2025 and worsen through 2026. These risks have already prompted some manufacturers to pressure suppliers to keep extra stock of power discrete components as a safeguard against disruption.
This situation highlights that the fundamental constraint lies less in wafer output and more in the back end, where limited investment in packaging and assembly continues to constrain availability.
Many automakers are now taking a proactive approach to semiconductor inventory. Some began stockpiling silicon MOSFETs and other discrete components as early as 2024. This defensive posture is particularly evident among automakers with major product launches in late 2025 and 2026 — such as Mercedes-Benz’s next-generation CLA.
Strategic takeaway: Balancing resilience and long-term risk
Memory vendors retain strong pricing power, while suppliers of silicon discretes and PMICs face ongoing allocation pressures. This imbalance between demand and supply is likely to persist through 2026 as AI data center demand continues to grow.
At the same time, automakers still have opportunities to diversify suppliers and selectively expand capacity, particularly in back-end packaging. The strategic challenge for OEMs will be balancing near-term resilience by bolstering their stocks with long-term risk management through broader supplier partnerships and alternative technologies.
Stay ahead of semiconductor shifts
Stay ahead of the curve with S&P Global Mobility’s Automotive E/E and Semiconductor Data and Forecasts on the AutoTechInsight platform, offering critical insights across electronic architectures and chip technologies. Planners and strategists rely on the E/E & Semiconductor Vehicle Domain to:
Track and forecast E/E and semiconductor technology growth.
Analyze OEM strategies and competitive positioning.
Understand customer purchasing and technology adoption paths.
Support decision-making with deep forecasts, supplier profiles, and component-level analytics.
Make informed, forward-looking decisions in the fast-changing automotive semiconductor landscape with trusted data and expert analysis from S&P Global Mobility.
This article was published by S&P Global Mobility and not by S&P Global Ratings, which is a separately managed division of S&P Global.