PulseAugur
EN
LIVE 18:26:44
Français(FR) Revolut a sorti Claude Code et Codex du quotidien de ~2 000 ingénieur·e·s. À la place, GLM auto-hébergé sur 56 B200 à 30 cents le million de tokens, dix fois mo

Revolut switches to self-hosted GLM on Nvidia B200s, cutting costs and improving speed

By PulseAugur Editorial · [1 sources] ·

Revolut has replaced its use of Claude Code and Codex with a self-hosted General Language Model (GLM) running on 56 Nvidia B200 GPUs. This new setup is ten times cheaper, costing $0.30 per million tokens, and has improved lead time by 70% by removing AI rationing for complex tasks. AI

IMPACT Demonstrates potential cost savings and performance gains by self-hosting LLMs on specialized hardware.

RANK_REASON Company internal product/infra change, not a frontier release or significant industry event.

Read on Mastodon — mastodon.social →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

Revolut switches to self-hosted GLM on Nvidia B200s, cutting costs and improving speed

How we ranked this

Signal score
4 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Tool
Company internal product/infra change, not a frontier release or significant industry event.
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
infra, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. Mastodon — mastodon.social TIER_1 Français(FR) · camilleroux ·

    Revolut has removed Claude Code and Codex from the daily use of ~2,000 engineers. Instead, self-hosted GLM on 56 B200s at 30 cents per million tokens, ten times mo

    Revolut a sorti Claude Code et Codex du quotidien de ~2 000 ingénieur·e·s. À la place, GLM auto-hébergé sur 56 B200 à 30 cents le million de tokens, dix fois moins cher. Le lead time a gagné 70 %, car l'IA n'est plus rationnée aux tickets difficiles. ⬇️ https:// twitter.com/guill…

    LINKS camilleroux.com/veille

RELATED ENTITIES

RELATED TOPICS