Tech — Page 36 | XenoSpectrum

Tech
Kioxia and Sandisk Announce World's Highest-Density 3D NAND Flash: 37 Gb/mm² Achieved with 332-Layer QLC
Kioxia and Sandisk achieved an areal density of 37 Gb/mm² with 332-layer QLC 3D NAND flash, overtaking Samsung's V10 (28 Gb/mm²), which leads in layer count, in terms of density. Through the CBA architecture, Toggle DDR6.0, and PI-LTT technology, they simultaneously improved density, speed, and power efficiency, delivering a design that meets the large-capacity storage demands of AI data centers.
August 5, 20268 min read
Tech
CFD to Revise Order Prices for GIGABYTE GPUs Starting in August: Impact on Retail Prices Still Unclear
CFD Sales has announced that it will revise order prices for GIGABYTE VGA products to 120–140% of current prices starting with orders placed on August 1. The notice also mentions the cancellation of backlogged orders, marking the first visible sign of rising GPU prices at the distribution stage, ahead of retail shelves.
August 5, 20266 min read
Tech
"Pass-ta-Key" Targets Google Synced Passkeys, Attacking Device Trust State Rather Than Cryptography
A cloud authenticator that Google designed to "protect passkeys from malware" was breached by an unprivileged malware attack unfolding across three stages. The cryptography was never broken. What broke was the management of trust in the operational layer of synchronization and re-enrollment.
August 5, 202612 min read
Tech
The Day Magnetic Tape Becomes the GPU's "Feeder": How Versity's GPUDirect Tape Gateway Bridges the Slow-Fast Divide
Versity has added RDMA support to its open-source S3 gateway, enabling direct transfers from tape archives to GPU memory. Amid soaring NVMe prices and server shortages, this low-cost setup aims for 20GBps using 50 parallel drives—how will it reshape storage design for AI training?
August 5, 20267 min read
Tech
16GB GPUs Overtake 8GB in Steam Survey, but the Gap Is Just 0.58 Points
In Valve's July 2026 Steam survey, 16GB VRAM reached 25.90%, edging out 8GB by 0.58 points to become the most common tier. Because participation is opt-in, the figures don't represent GPU performance rankings or overall market adoption rates.
August 5, 20265 min read
Tech
How Kiro Crew Hands Off Development Work: AI Agent Operations Transformed by Cron and Webhooks
Built on Kiro CLI, Kiro Crew is an open-source workspace that combines memory, scheduling, and multiple agents to keep work moving even while developers are away. The software itself is free, but inference draws on Kiro's credits.
August 5, 20265 min read
Tech
NVIDIA Open-Sources cuFile API, Ushering in an Era Where GPUs Drive Storage Directly
NVIDIA has open-sourced cuFile, its API for direct GPU storage access, and is driving standardization of the SCADA architecture—where GPUs control storage I/O without going through the CPU—together with more than 40 companies. As the bottleneck in AI inference shifts from compute to data delivery, structural change is beginning across the storage industry.
August 5, 20269 min read
Tech
The Night OpenAI Dragged Apple's Lawsuit Into the Court of Public Opinion
OpenAI responded to Apple's trade secrets lawsuit not in court but in a blog post. It published attorney-to-attorney emails verbatim, working to dismantle Apple's narrative of a "misaddressed email" and "broken-off contact." Ahead of a ruling on the injunction, the aim is to buy time in the court of public opinion.
August 4, 202611 min read
Tech
B200 at $5.66 an Hour: Why Rising GPU Rental Index Numbers Don't Equal Longer Depreciation
Silicon Data's GPU rental index shows B200 climbing to $5.66 per hour, while the same H100 splits into different tiers depending on provider. Here's the framework for comparing hourly rates without conflating them with equipment lifespans or depreciation costs.
August 4, 20266 min read
Tech
Lexar Launches DDR5 Memory Kit Using CXMT DRAM: The Reality That New Chinese Memory Suppliers Don't Immediately Lower Prices
Lexar has listed a 32GB DDR5-7200 kit using domestic Chinese DRAM for 3,999 yuan on JD.com. While clock speeds have advanced, the emergence of a new supplier doesn't necessarily translate into cheaper memory.
August 4, 20265 min read