Chinese Workshops Dismantle RTX 5090 GPUs for AI Servers
Chinese workshops dismantle GeForce RTX 5090 cards to harvest GB202 silicon and GDDR7 memory, creating dual-slot blower GPUs for dense…
Coverage desk
Tag: Silicon
Current desk
Newest reporting
Musk Targets Mid-2027 Test Chip for Texas Terafab Facility
Elon Musk sets a mid-2027 test chip goal for Terafab in Texas, but semiconductor experts question whether the joint Tesla-SpaceX…
AMD Zen 6 APU Surfaces with 24GB LPDDR5 in Benchmarks
An unreleased AMD Zen 6 APU with 24GB of LPDDR5-9600 memory has appeared in OpenBenchmarking tests, hinting at next-gen handheld…
How Visko Deploys Real-Time Video World Models on Cloud TPUs
Visko deployed its Orbis live video world model on a single Cloud TPU slice, hitting 1.07x playback speed via JAX,…
Applied Materials Opens $5B EPIC Center for AI Chip Research
Applied Materials opened its $5B EPIC Center in Silicon Valley, offering shared cleanrooms to accelerate collaborative semiconductor process innovation.
Intel Details GPU D270 Accelerator for Data Center Inference
Intel branded its Crescent Island AI accelerator as the GPU D-Series, revealing the air-cooled GPU D270 card with 160GB LPDDR5X…
Google Launches Project Suncatcher to Test AI TPUs in Orbit
Google launched Project Suncatcher to test four TPUs in orbit aboard a SpaceX rocket, evaluating radiative cooling, solar power, and…
Efficient Computer Closes $97M Round for Dataflow Chips
Efficient Computer raised $97 million in Series B funding to scale its Electron E1 spatial dataflow processor for edge robotics…
AMD EPYC Venice Benchmarks Target Nvidia Vera: A Technical Look
AMD reveals new EPYC Venice server benchmarks against Nvidia Vera, showing big throughput gains while mixing GCC compiler versions and…
AI Inference Demands Reshape Datacenter Memory and Silicon
At the AI Infra Summit, chipmakers and cloud hyperscalers tackled the memory wall and rising token costs to scale AI…