Goldman: Apple DRAM Buy Won't Hurt Margins

🔑 Enhanced Key Takeaways

• •Apple's DRAM procurement strategy is heavily influenced by the integration of on-device generative AI models, which require significantly higher memory bandwidth and capacity compared to previous iPhone generations.
• •The 2.4 EB (Exabyte) figure represents a strategic shift toward standardizing 16GB and 24GB RAM configurations across the 2026 iPhone lineup to support complex local inference tasks.
• •Goldman Sachs' analysis suggests Apple is leveraging its massive cash reserves and long-term supply agreements to lock in favorable pricing, effectively insulating itself from the cyclical volatility of the broader memory market.

🛠️ Technical Deep Dive

• •Transition to LPDDR5X: The 2026 iPhone lineup utilizes LPDDR5X, offering higher data rates (up to 9600 Mbps) compared to standard LPDDR5 to handle the increased memory bandwidth requirements of local AI processing.
• •Memory Architecture: Apple is implementing a unified memory architecture that allows the Neural Engine and GPU to access the same high-speed memory pool, reducing latency for real-time AI inference.
• •Capacity Scaling: The move to 2.4 EB total capacity is driven by the need to keep large language models (LLMs) resident in memory to avoid the latency penalties of swapping to NAND flash storage.

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline