Tag: ai compute
-
From compute to capability
The throughline of recent AI progress has been scale: larger systems trained on more data with more computation. Understanding that loop helps explain where the market is heading. Read more
-
AI’s growing energy demand
As data centres expand, their share of electricity demand is rising. Energy availability is becoming a genuine constraint on how fast AI capacity can scale. Read more
-
The data-centre construction boom
Monthly spending on data-centre construction has surged as companies race to add capacity for AI workloads — a build-out with real consequences for cost, power and timelines. Read more
-
Why training compute keeps growing
The computation used to train leading AI systems has risen exponentially, roughly doubling on a short cadence. More compute has consistently translated into more capable models. Read more