Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-03-12 · source-backed
Four chips detailed: MTIA 300 (in production), 400 (lab testing), 450 and 500 (GenAI inference, 2027). From 300 to 500: HBM bandwidth up 4.5x, compute FLOPS up 25x. New chips every six months. The most aggressive custom silicon roadmap from any hyperscaler, reducing NVIDIA/AMD dependency. Meta Engineering
Each link below shares sources, entities, or timing with this story.
NVIDIA uses TSMC / Shared entities / Shared topic / What happened next / Tension
Linked by a graph relationship (NVIDIA uses TSMC); both cover Meta, Nvidia; overlapping topics (chip, custom, dependency, meta).
AMD partners with Meta / Shared entities / Shared topic / What happened next
Linked by a graph relationship (AMD partners with Meta); both cover Meta, Nvidia; overlapping topics (chip, compute, custom, inference, meta).
Anthropic uses NVIDIA / Shared entities / Shared topic / What happened next
Linked by a graph relationship (Anthropic uses NVIDIA); both cover AMD, Meta, NVIDIA; overlapping topics (inference, meta).
NVIDIA invested in OpenAI / Shared entities / What happened next / Tension
Linked by a graph relationship (NVIDIA invested in OpenAI); both cover AMD, Meta, Nvidia; picks up the AMD thread on 2026-07-25.
NVIDIA released DGX Spark / Shared entities / Shared topic / What happened next / Tension
Linked by a graph relationship (NVIDIA released DGX Spark); both cover AMD, NVIDIA; overlapping topics (bandwidth, generation).
Anthropic uses NVIDIA / Shared entities / Shared topic / What happened next
Linked by a graph relationship (Anthropic uses NVIDIA); both cover Meta, Nvidia; overlapping topics (chip, custom, silicon).
Anthropic uses NVIDIA / Shared entities / What happened next
Linked by a graph relationship (Anthropic uses NVIDIA); both cover AMD, Meta, NVIDIA; picks up the AMD thread on 2026-07-27.
Jensen Huang works at NVIDIA / Shared entities / Shared topic / What happened next
Linked by a graph relationship (Jensen Huang works at NVIDIA); both cover Meta, NVIDIA; overlapping topics (bandwidth, custom).