Unified memory becomes edge AI’s breaking point
Why memory architecture now decides AI capability.
The AIP zoo
180 creatures vying for your attention.
AMD buys Taalas to sharpen inference
Toronto chip specialist joins AMD roadmap.
Tensordyne bets on math to beat Nvidia
Start-up relies on logarithmic math to replace multiplication functions with addition wherever possible, and aims to cut power and gain room for memory.
Efficient Computer pulls $60M to push energy-efficient AI to the edge
Pittsburgh start-up rewrites chip architecture for energy-starved AI.
Agentic AI recasts host CPUs
Memory and orchestration drive procurement.
AI chip scarcity reshapes infrastructure strategy
Custom silicon moves into deployment planning.
Lumai’s free-space bet on optical inference
Lumai’s head of product explains the company’s approach to optical AI compute.
Asygn’s Colibry NPU tries to make edge AI live on scraps of power
Asygn wants Colibry to bring vision AI to the smallest devices.
Netrasemi announces its 6th chip
Six-year-old Indian AI processor company stays at 12 nm.
Intel aims Crescent Island at inference
LPDDR5X memory targets practical AI deployment.
Neurophos: Interview with CEO Patrick Bowen
Understanding the math needed for AI is critical.