{"slug":"revolut-makes-self-hosted-glm-models-its-default-for-ai-coding-across-roughly-2-000-engineers-on-56-nvidia-b200-gpus","url":"https://findausecase.com/use-cases/revolut-makes-self-hosted-glm-models-its-default-for-ai-coding-across-roughly-2-000-engineers-on-56-nvidia-b200-gpus","title":"Revolut makes self-hosted GLM models its default for AI coding across roughly 2,000 engineers on 56 Nvidia B200 GPUs","description":"Revolut rolled out AI coding tools to roughly 2,000 engineers and moved from commercial services to self-hosted GLM models served on 56 Nvidia B200 GPUs, with a model router sending vision queries to Gemini or GPT. The company reports roughly 90 percent lower inference costs, a 60 percent improvement in coding time and a 70 percent improvement in lead time.","company":"Revolut","industry":"Financial Services","country":"United Kingdom","aiCapabilities":["Generative AI"],"technology":["GLM-5.3","GLM-5.1","Nvidia B200","Gemini","GPT"],"deployment":"Unknown","problemStatement":"Revolut was skeptical about the cost of AI coding tools, seeing engineers spend $1,000 to $10,000 a month with little return, so it launched them with strict budget limits that left engineers with only a limited number of tokens.","solutionApproach":"Revolut rolled out AI coding tools to roughly 2,000 engineers and made self-hosted GLM models its default, initially GLM-5.1 and then GLM-5.3, served on 56 Nvidia B200 GPUs. A model router sends vision queries to models such as Gemini or GPT while code generation stays on GLM.","businessValue":"Revolut says it cut AI coding inference costs by roughly 90 percent in less than a year, and reports a 60 percent improvement in coding time and a 70 percent improvement in lead time following the rollout. Self-hosting GLM 5.1 at $0.30 per million tokens cost about a tenth of using OpenAI's Codex or Anthropic's Claude Code.","evidence":{"band":"high"},"sourceUrl":"https://www.theregister.com/a/5307345","dates":{"publishedAt":"2026-11-12T05:46:43.462Z","publishedAtSource":"pipeline","updatedAt":"2026-11-12T05:46:43.462Z"},"license":"Open for reading and citing with a link to https://findausecase.com/use-cases/revolut-makes-self-hosted-glm-models-its-default-for-ai-coding-across-roughly-2-000-engineers-on-56-nvidia-b200-gpus. Bulk republication requires permission."}