
AWS has released Graviton5, a new processor built from four smaller chiplets rather than one large chip, containing 192 cores designed to handle databases and agentic AI applications that require low latency. The Graviton5 delivers 2.4 times more performance per socket than its predecessor while using a more power-intensive 3 nanometer manufacturing process, making it faster but hotter than previous generations. New M9g and M9gd instances based on Graviton5 offer 31.9 to 33.6 percent better price-to-performance compared to equivalent older models, though they have less memory capacity because memory is expensive relative to compute. The design tradeoff prioritizes responsiveness for systems that need low latency over power efficiency, supported by faster DDR5 memory and technology for extending memory capacity across the system.

South Korea’s government and top tech companies are committing $1 trillion to several flagship megaprojects that could bolster global memory chip supply, build new AI data centers and spur commercial deployment of humanoid robots by 2028. The announcement comes as South Korean companies such as Samsung and SK Hynix have enjoyed record profits and stock valuations due to the AI industry’s demand for memory chips—with the subsequent supply strain leading to memory chip shortages and higher prices

The world's two largest memory chip companies vow to build more memory lab fabs as South Korea positions itself as an AI tech powerhouse country.

FERC filings show AI developers and grid operators converging on stricter readiness rules to separate real power demand from speculative projects.
Want to go deeper than the news? Explore live, cohort-based AI courses taught by practitioners.
Browse AI courses on Maven