The industry everything else quietly runs on. I cover it from the fab up, process nodes and lithography and the foundry wars, plus the export controls now deciding who even gets to build what. This is the manufacturing and geopolitics side of silicon, kept deliberately apart from the chip-and-platform coverage.
Nvidia's RTX Spark brings 20 ARM cores and 6,144 CUDA cores to Windows laptops this fall. The full architecture breakdown, Geekbench comparisons against Apple…
NVIDIA's B300 yields are stabilizing. That is not the story. The story is CoWoS: TSMC's packaging bottleneck that sits between Nvidia and the demand…
Qualcomm expands compute architecture to desktops with the X2 Elite, splits mobile silicon into standard and Pro tiers, and pushes toward a $6B automotive…
Qualcomm recently launched the Snapdragon 6 Gen 5 and Snapdragon 4 Gen 5, focusing on a feature called Smooth Motion UI, which optimizes chip…
CATL Qilin vs BYD Blade compared across chemistry, cooling, energy density, safety, and market share in this 2026 technical breakdown.
Jetson Thor owns the humanoid robot; Qualcomm Dragonwing is fighting for the high-volume middle. The 2026 edge robotics compute split, broken down.
Mistral AI sells intelligence through models and APIs, Cerebras sells raw compute through wafer-scale silicon. A no-nonsense breakdown of two very different AI revenue…
Qualcomm's Dragonwing IQ10 packs an 18-core Oryon CPU, up to 700 TOPS, a safety island, and dense sensor I/O. A voice-driven architecture deep dive,…
The USB Power Delivery (PD) 3.1 introduced a significant advancement by increasing power capacity from 100W to 240W through new voltage levels. This change…
The post details the rapid advancement of AI from 2023 to 2026, highlighting significant investments, partnerships, and innovations in infrastructure, robotics, and consumer devices,…
In 2026, storage prices are rising due to DRAM and NAND shortages, pushing some companies toward failure. Most brands are assemblers rather than manufacturers,…
Renesas, Microchip and Ceva are pushing AI inference onto microcontrollers at milliwatt budgets. Inside the TinyML architecture war for the volume edge.
All LoRa chips are produced by Semtech, which acquired Cycleo's technology and established the LoRaWAN protocol. This ecosystem, primarily driven by utilities and smart…
The evolution of edge AI processing from the Myriad 2's 100 gigaflops to the Dragonwing Q-8750's 77 TOPS illustrates a shift towards powerful, local…
The AI revolution began in 2015 with foundational developments like TensorFlow and the transformer model. By 2022, ChatGPT brought widespread attention to AI, leading…
One maker's tech digest, weekly. Chips, AI, gaming, EVs — no spam.