{"slug": "at-t-routes-40-of-ai-workloads-through-open-weight-models-cutting-coding-costs", "title": "AT&T routes 40% of AI workloads through open-weight models, cutting coding costs by 56%", "summary": "AT&T now routes roughly 40% of its internal AI requests through open-weight models, cutting AI coding costs by 56% while absorbing only a 2% decline in output quality, the company said. The telecom giant processes approximately 45 billion tokens per day, up from around 8 billion a year earlier, and targets 60-70% open-model usage within the next year using an intelligent routing system built on LiteLLM and a custom AI gateway. AT&T also launched OTel 2.0, a post-trained open-weight model for telecom data developed with the GSMA's Open Telco AI initiative, trained on over 400 billion tokens and built on AMD GPUs and Microsoft Foundry.", "body_md": "# AT&T routes 40% of AI workloads through open-weight models, cutting coding costs by 56%\n\nThe telecom giant is processing 45 billion AI tokens daily and plans to push open-model usage to 70% as it rethinks spending on closed alternatives from OpenAI and Anthropic.\n\nAT&T is quietly pulling off one of the more aggressive enterprise AI pivots in recent memory. The telecom giant now routes roughly 40% of its internal AI requests through open-weight models, and it’s not slowing down. The company has set a target of 60-70% within the next year.\n\nThe math behind the decision is hard to argue with. AT&T has slashed AI coding costs by 56% while absorbing only a 2% decline in output quality. For certain complex workloads, the savings are even more dramatic, with cost reductions hitting 80-90% compared to closed-model alternatives.\n\n## The tokenomics of telecom-scale AI\n\nAT&T’s AI consumption has grown at a pace that would make any CFO nervous. The company now processes approximately 45 billion tokens per day, up from around 8 billion just a year ago. That’s roughly a 5.6x increase in twelve months.\n\nAT&T calls its approach “tokenomics,” and the logic is straightforward: not every AI request needs the most expensive model in the room. The company uses an intelligent routing system built on LiteLLM and a custom AI gateway to match tasks with the appropriate model. Simpler requests get directed to open-weight options like [Nvidia](https://cryptobriefing.com/markets/nvidia/) Nemotron, [Meta](https://cryptobriefing.com/markets/meta/) Llama, and [Google](https://cryptobriefing.com/markets/alphabet/) Gemma. More complex tasks can still tap into closed models when the quality threshold demands it.\n\n## Open models closing the gap\n\nAT&T executives have stated that open-weight models have narrowed the performance gap with proprietary alternatives, which has prompted the company to rethink its investments in offerings from [Anthropic](https://cryptobriefing.com/markets/anthropic/) and [OpenAI](https://cryptobriefing.com/markets/openai/).\n\n### AI, tech, and the markets they move—in one daily briefing.\n\nDaily. Free. Join 34,000+ readers across crypto, finance, and policy.\n\nAT&T isn’t just consuming open models. It’s building them. The company launched OTel 2.0, a post-trained open-weight model specifically designed for telecom data. The model was developed in association with the GSMA’s Open Telco AI initiative, trained on over 400 billion tokens, and built using [AMD](https://cryptobriefing.com/markets/amd/) GPUs and [Microsoft](https://cryptobriefing.com/markets/microsoft/) Foundry.\n\n## The broader enterprise shift\n\nAT&T isn’t operating in a vacuum. The move toward open-weight models reflects a growing trend across enterprise AI adoption, driven by three concerns: cost efficiency, customization, and data sovereignty.\n\nThe company is also keeping an eye on Chinese open-weight models, though it hasn’t adopted any. Security and compliance concerns have kept those options on the evaluation bench rather than in production.\n\n**Disclosure:** This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our\n\n[Editorial Policy](https://cryptobriefing.com/editorial-policy/).", "url": "https://wpnews.pro/news/at-t-routes-40-of-ai-workloads-through-open-weight-models-cutting-coding-costs", "canonical_source": "https://cryptobriefing.com/att-open-weight-ai-models-cost-savings/", "published_at": "2026-09-29 17:37:40+00:00", "updated_at": "2026-09-29 17:47:17.106868+00:00", "lang": "en", "topics": ["ai-infrastructure", "large-language-models", "ai-products", "mlops", "ai-chips"], "entities": ["AT&T", "OpenAI", "Anthropic", "Nvidia", "Meta", "Google", "LiteLLM", "GSMA"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/at-t-routes-40-of-ai-workloads-through-open-weight-models-cutting-coding-costs", "markdown": "https://wpnews.pro/news/at-t-routes-40-of-ai-workloads-through-open-weight-models-cutting-coding-costs.md", "text": "https://wpnews.pro/news/at-t-routes-40-of-ai-workloads-through-open-weight-models-cutting-coding-costs.txt", "jsonld": "https://wpnews.pro/news/at-t-routes-40-of-ai-workloads-through-open-weight-models-cutting-coding-costs.jsonld"}}