# Microsoft — AT&T and Microsoft scale trillion-token workloads with Microsoft Foundry and AMD

- Company: Microsoft (microsoft.com)
- Announced: 2026-07-23T18:30:00+00:00
- Category: infrastructure-release
- Subject: Azure AI
- Models affected: Phi-4, OSS-120B, Gemma-4
- Source: https://azure.microsoft.com/en-us/blog/att-and-microsoft-scale-trillion-token-workloads-with-microsoft-foundry-and-amd/
- Record: https://forck.live/items/2446-at-t-and-microsoft-scale-trillion-token-workloads-with-microsoft-foundry-and-amd

Microsoft announces that AT&T used Microsoft Foundry Managed Compute with AMD Instinct MI300X GPUs to train its OTel2.0 telecom AI model, processing about 1 trillion tokens and saving tens of millions of dollars by using open models like Phi-4 instead of frontier models.

## Evidence

Verbatim from https://azure.microsoft.com/en-us/blog/att-and-microsoft-scale-trillion-token-workloads-with-microsoft-foundry-and-amd/:

> AT&T processed approximately 1T tokens, consisting of raw documents from GSMA supplemented by synthetic data generated. Generating the data using open-source models like Phi-4, served by Microsoft’s Foundry Managed Compute, saved tens of millions of dollars versus using frontier models.

---

Record: https://forck.live/items/2446-at-t-and-microsoft-scale-trillion-token-workloads-with-microsoft-foundry-and-amd
Catalogue: https://forck.live/llms.txt
Feed: https://forck.live/feed.md