[NVIDIA Vera Enhances AI-Native Storage Processing Efficiency, Covering Key Areas like Encryption, Compression, and Data Recovery] DeepTechFlow reports that on August 5, NVIDIA released benchmark results for the Vera BlueField-4 STX storage processor, claiming it can accelerate critical data path processing tasks in AI-native storage scenarios, such as encryption and decryption, Reed-Solomon recovery, CRC32C integrity checks, as well as compression and decompression. The article shows that Vera achieved higher throughput performance in multiple tests compared to x86 processors, with notable improvements in recovery, integrity checks, and compression-related tasks. The report highlights that as AI workloads increase in concurrency and context scale, storage systems must handle not only basic read and write operations but also CPU-side tasks like data protection, verification, compression, and recovery. NVIDIA believes that Vera, with its superior single-core performance, memory bandwidth, and multithreading capabilities, can enhance storage processing efficiency within limited power and thermal budgets, enabling AI-native storage platforms to support higher concurrent data streams and service density.