Micron recently signaled a significant shift in memory requirements for the automotive industry.
The company stated that Level-4 autonomous vehicles will need over 300GB of DRAM, a massive leap from the roughly 16GB found in today's average car. This nearly 19-fold increase highlights how self-driving technology is set to become a major driver of memory demand, potentially consuming vast amounts of supply as it scales.
This forecast arrives amid a severe global memory shortage. First, the 'scarcity narrative' is powerful, as Micron itself has warned it can only supply about half to two-thirds of its customers' needs through 2026. Adding a huge new source of demand like autonomous vehicles will likely intensify competition for limited supply and support strong pricing for memory makers.
Second, this 300GB figure is the next logical step on an established path. High-end computing platforms for vehicles, like NVIDIA's DRIVE Thor, are already shipping with 128GB of LPDDR5X RAM. This establishes a high-performance baseline, making Micron's projection for full autonomy seem credible as sensor data and AI models become more complex.
So, why isn't a newer, potentially cheaper technology like High Bandwidth Flash (HBF) the answer? Micron's caution is rooted in HBF's intended role and timeline. HBF is being developed as a high-capacity tier between ultra-fast HBM and slower SSDs, mainly for AI inference, not for the real-time, low-latency processing that autonomous driving safety systems demand. With proponents targeting significant adoption around 2030, DRAM remains the primary, proven solution for the near future.
In essence, Micron's announcement solidifies DRAM's critical role in the next generation of vehicles, placing the automotive sector at the center of the ongoing memory supply crunch.
- Glossary -
- L4 Autonomous Driving: Level 4, or high automation, where the vehicle can handle all driving situations within a specific operational design domain (e.g., a geofenced urban area) without human intervention.
- DRAM (Dynamic Random-Access Memory): A type of volatile semiconductor memory that is essential for the rapid data processing required by computers, servers, and increasingly, advanced vehicles.
- HBF (High Bandwidth Flash): An emerging memory standard that uses NAND flash technology to offer higher capacity than HBM at a lower cost, positioned as a middle tier for AI inference workloads rather than a direct DRAM replacement.