🤖 AI Summary
OVHcloud CEO Octave Klaba warned that cloud prices could rise about 5–10% by mid-2026 (or sooner) as AI-driven demand pushes up the cost of key hardware like RAM and NVMe storage. He pointed to industry data (TrendForce) showing sharp memory price inflation—DDR4 up ~158% and DDR5 up ~307% since Sept 2025—and vendor moves such as Samsung raising some prices by ~60%. Klaba said stockpiling can delay the passthrough for 6–12 months, but expects server costs for OVH to climb 15–25% in the April–September 2026 window; he caveats the forecast is based on November 2025 trends and could accelerate.
For the AI/ML community this matters because memory and NVMe are direct drivers of training and inference costs for large models and data pipelines. Higher hardware prices will translate into higher operating expenses for cloud-based model training, larger batch/throughput jobs, and storage-heavy workflows—pushing teams toward efficiency measures (quantization, sparsity, distillation, memory-efficient architectures), cost-focused scheduling (spot instances, regional optimization, batching), and hybrid or on-prem strategies where feasible. Providers may absorb some increases short-term, but expect tighter margins or customer price adjustments, reinforcing incentives to optimize models and infrastructure choices as hardware-driven unit costs rise.
Loading comments...
login to comment
loading comments...
no comments yet