A new GPU generation is arriving. Plan the migration of a live 2,048-GPU cluster.

Plan a GPU-generation migration around usable workload capacity, site readiness, compatibility and a tested rollback path.

Updated Sep 2026 · Learn how AI infrastructure works through explanations, worked examples and diagrams, then practise applying it to interview questions.

Plan a GPU-generation migration around usable workload capacity, site readiness, compatibility and a tested rollback path.

The full answer is part of Premium. A free account includes more answers in each topic, but does not unlock this one.

20 answers per topic instead of 10, plus saved progress and bookmarks · no cardor unlock all 316 remaining answers · ₹2,000 / $25

The concepts behind this question

Use these concept explanations to review the ideas behind this question.

Foundational

🧮 Open Weights & Serving Engines

Capacity Planning for Open-Weights FleetsMatch forecast traffic to a measured replica operating envelope, then check failure domains, rollout capacity, readiness time and cost. Output tokens alone omit prompt work and live state; total parameter count alone does not determine a replica's throughput.

Foundational

🖧 Hardware & Cluster Build-Out

Colocation, Power Contracts and Site SelectionA suitable site must deliver the required rack power, cooling and service capacity. Contract terms and facility lead times shape both cost and deployment dates.

Advanced

🧮 Napkin Math & Capacity🔒 Premium

Power and Datacenter ConstraintsGPU capacity also needs electrical and cooling capacity. Build the facility estimate from complete system loads, rack density and the site’s operating limits.

Advanced

🚀 Inference & Serving🔒 Premium

Reasoning Budgets and Answer LatencyReasoning tokens consume generation time and retained state before an answer becomes usable. Separate the first generated token from the first answer token, reserve output capacity, and evaluate reasoning limits against task success.

Explore the full curriculum →

UP NEXT ON YOUR JOURNEY

Next in this trackA vendor claims their accelerator beats an H100 at half the price. How do you evaluate that?Next in this trackDesign the health check that decides whether a node may accept jobs.Next in this trackA node fails burn-in twice and passes on the third attempt. Do you accept it?

FEDITOR'S NOTE

Compare capacity by workload, not device count alone. Keep old capacity until replacement performance and recovery are accepted.

DISCUSSION · 0

No comments yet — be the first to share your approach.