↗
Nvidia Puts $3.5B Behind MediaTek AI Infrastructure
Nvidia invests $3.5 billion in MediaTek as the companies expand NVLink Fusion, custom AI accelerators, Spark PCs, and vehicle platforms.
CH CHANNEL ARCHIVE
Tag: AI
Artificial intelligence models, systems, and products.
01 INCOMING
ARCHIVE
↗
↗
Google Antigravity Puts Gemini Agents to Work in Teams
Google says Antigravity's Gemini 3.7 Flash agent teams solved research problems, built a RISC-V simulator, and improved open-source code.
↗
SpaceX Targets Gas Turbine Bottleneck for AI Power
SpaceX says an in-house turbine foundry could shorten AI power waits, exposing the manufacturing bottleneck behind new data-center capacity.
↗
Pollen Robotics Launches $399 Microduck Biped
Pollen Robotics' 25-centimeter Microduck has 15 motors, open-source control software, and a $399 preorder aimed at physical-AI developers.
↗
a16z Raises $1.1B for AI Hardware
Andreessen Horowitz has raised $1.1 billion for a fund targeting the chips, memory, networking, power, and robots behind AI infrastructure.
↗
NVIDIA Jetson Orin Nano 2 Brings Edge AI to Robots
NVIDIA announced Jetson Orin Nano 2, an entry level edge AI computer promising 2x the inference of its predecessor for robots, drones, and vision systems.
↗
Intel Outlines Three Architectures for Agentic AI
At Hot Chips 2026 Intel detailed Diamond Rapids, Crescent Island, and Wildcat Lake, three architectures that scale agentic AI from data center to edge.
↗
Intel Panther Lake CPUs: Which One for AI and Gaming?
Intel's Panther Lake X7 358H and Arc B390 bring 50 TOPS of NPU power and strong integrated graphics to laptops without a discrete GPU.
↗
IBM Builds a Mainframe Core That Speaks Arm and Z
IBM revealed a dual-architecture mainframe processor at Hot Chips 2026 whose cores run Arm and IBM Z code natively, for the next Z and LinuxONE generation.
↗
What Is a TPU? TPU vs GPU, Explained
A TPU is Google's custom AI accelerator chip. What a TPU does, how TPUs vs GPUs compare for AI training and inference, and who builds them.