TickerTain
TickerTain
NewsroomShortsPortfolioConvergence
NewsroomShortsPortfolioConvergence
←
▶ 6:04 · $CBRS · AWS partners with Cerebras for disaggregated prefill-decoding inference architecture
episode briefing
SemiAnalysis

AWS Trainium: How Amazon Built Their Own AI Chips | Researcher Conversations at GTC

2026-04-30 · 3 company · 3 thematic
sentiment
3 bull0 bear0 neu
speakers
randa

Leads product marketing for AI infrastructure at AWS, discussing the company's 15-year Nvidia partnership, 2M GPU fleet expansion, Trainium 3 custom silicon roadmap, and Cerebras disaggregated inference partnership.

quote
“recently the announcement of the the Cerebras and uh and the Trainium uh product with uh with supporting disaggregated inference. Yeah. And the team is also similar there, like bringing down the cost for our customers, especially for infer…”
—randa
now playing · $CBRS
$NVDA···bullish· highranda
AWS to add 1M Nvidia GPUs in 2026, deepening 15-year partnership
AWS is expanding its Nvidia GPU fleet by 50% in a single year, signaling massive sustained demand for Nvidia hardware and a deepening strategic partnership that includes upcoming…
$AMZN···bullish· high· posranda
AWS Trainium 3 delivers 30-40% better price-performance, scaling to 1M+ chips in 2026
Amazon's custom Trainium 3 chips offer 2-3x performance over prior generation and 30-40% better price-performance vs alternative accelerators, with plans to deploy 1M+ chips in 20…
$CBRS···bullish· mediumranda
AWS partners with Cerebras for disaggregated prefill-decoding inference architecture
AWS and Cerebras are collaborating on a disaggregated inference architecture separating prefill and decode stages to drive down dollar-per-token costs for customers deploying GenA…