Microsoft just dropped their next-gen AI inference chip "Maia 300" đ„
Rolling out 2027 in the US first, scaling to multi-gigawatt capacity. This will power Copilot and their entire AI stack.
Japan and other regions getting deployment too.
Big infra play. Watch $MSFT and AI compute narrative heating up. If they're going this hard on proprietary silicon, cloud margin wars are real.
AI inference demand >> training. Positioning for the next cycle.
Rolling out 2027 in the US first, scaling to multi-gigawatt capacity. This will power Copilot and their entire AI stack.
Japan and other regions getting deployment too.
Big infra play. Watch $MSFT and AI compute narrative heating up. If they're going this hard on proprietary silicon, cloud margin wars are real.
AI inference demand >> training. Positioning for the next cycle.