topic
AI compute
4 dispatches on AI compute.
Samsung's Memory Stack Goes Vertical: What zHBM, zNAND-O, and V10 BV-NAND Actually Mean
Samsung unveiled three wafer-bonded memory architectures at FMS 2026—zHBM, zNAND-O, and V10 BV-NAND. Here's what the specs say, what's real today, and what's still a concept.
27,500 GPUs for Japan: NVIDIA's Vera Rubin Bet Is About Cost Per Token, Not Peak Performance
NVIDIA's Vera Rubin AI factory in Japan—27,500 Rubin GPUs, 140 megawatts, backed by the Japanese government—signals that the real AI arms race is now about cost-per-token infrastructure, not raw benchmark scores.
NVIDIA's RTX Spark Is NVIDIA's First PC Chip — and It's Gunning for Apple Silicon's Crown
NVIDIA unveiled the RTX Spark Superchip at Computex 2026: a Grace CPU plus Blackwell GPU on a single Windows-on-Arm SoC, backed by every major OEM, shipping this fall. Here's what the platform actually means.
NVIDIA Wants to Own the Whole PC Now
At Computex 2026, NVIDIA unveiled RTX Spark — a Blackwell GPU fused with a MediaTek-designed CPU and unified memory. It's NVIDIA's first real play for the consumer PC stack, and the real competition isn't Apple. It's the idea that the PC platform should have an owner at all.