Amazon BedrockBedrock vs Vertex AI vs Foundry: Your Cloud Commit Decides
Claude runs on Bedrock, Vertex AI and Foundry at the same list price, so the committed-spend agreement you already signed decides. Bedrock is the default, Vertex is the only Gemini home, and Foundry is the pick for GPT but the weakest for Claude.
October 25, 2026 · 14 min readprovisioned throughputAnnual Azure and Google AI reservations break even at 84% and 75%
Azure and Google now meter reserved model capacity at pay-as-you-go token weights, so a monthly term needs about 99% utilization to save money. Annual terms break even at 84% on Azure and 75% on Google; Bedrock publishes too little to compute.
October 22, 2026 · 15 min readfine-tuning costFine-Tuning vs RAG Cost: The Training Bill Isn't the Bill
Training a custom model costs about $30. Serving the same weights costs $326 to $28,908 a month depending on the platform. The fine-tuning-versus-RAG decision is a serving-contract decision, and the break-even is 683,000 queries.
September 9, 2026 · 20 min readEquinixEquinix Set a Date, Not a Price. Cap Your Commit Term.
Equinix Inference Exchange arrives Q1 2027 with 200+ open models, two tenancy modes and no disclosed price. Your token-API commitments already clear that date; your GPU and colo deals do not. Cap the term or attach a re-price trigger.
September 3, 2026 · 11 min readAI capacity procurementMicrosoft's GPUs Sit Unplugged. Buy Delivery, Not Capex.
A Guardian investigation put a measured number on the gap between Microsoft's announced AI capacity and its installed accelerators. Azure's own documentation already says a PTU reservation guarantees you a price, not capacity — which is why a capacity clause has to name region, SKU, quantity, delivery date and remedy.
August 17, 2026 · 13 min read