AMD Acquires Taalas to Embed AI Model Weights Directly in Silicon
The Canadian startup's approach bakes neural network parameters onto chips, bypassing traditional high-bandwidth memory as inference workloads evolve beyond GPU-centric architectures.
Every BriefAsia story tagged Amd, newest first.
The Canadian startup's approach bakes neural network parameters onto chips, bypassing traditional high-bandwidth memory as inference workloads evolve beyond GPU-centric architectures.

The company's second-quarter results underscore its strategy to challenge Nvidia by mirroring the competitive tactics that won it ground in data center processors.

Taiwan's packaging specialist targets June launch for panel-level technology, posts strongest quarterly profit in over two years on AI memory demand

New AI-native development suite promises 3.3x inference performance gains and automated optimization for enterprise AI deployment

Industrial computing veteran extends AI infrastructure from data center to edge with latest AMD processor lineup

Production begins on the 72-accelerator system designed to challenge Nvidia's data center dominance, with shipments starting Q3 2026

Chinese cloud providers lock in long-term processor contracts as Intel and AMD shipments tighten, marking a shift from once-stable supply chains to semiconductor-style allocation.