OpenAI Unveils Custom Jalapeño AI Inference Chip, Samsung Reportedly Supplying HBM4

OpenAI has introduced its first custom AI inference chip, named Jalapeño, just before NVIDIA's latest earnings report. Initial benchmarks suggest the chip offers significantly higher throughput per kilowatt and lower end-to-end latency compared to NVIDIA's GB200 and GB300 rack systems. Samsung Electronics is reportedly supplying the HBM4 memory for Jalapeño, which could intensify the competition for HBM4 supply in the semiconductor market.