Enterprise AI data pipeline bottleneck converts GPU capex into all-flash storage demand branch condition conversion continues enterprises defer GPU cluster · AI 1 GPU compute added enterprises · neo-cloud buyers training + inference fleets 2 · data pipeline jams — GPUs waiting for data 3 · all-flash NAS/SAN fabric local flash · data platform for AI 4 · revenue conversion +18% YoY · $1.216B · ~62% of total revenue · Q4 FY26 AI wins ~1,100 · prev ~400 capacity tiering bypassed for AI 5 · pivotal branch does AI storage spend hold? durable tailwind AI workloads keep converting AFX + NVIDIA co-engineered reversion to base enterprises defer · pipeline stalls mid-single-digit growth resumes