Why are telecom operators betting their AI strategy on open models?

Telecom operators are shifting their AI strategies toward open models. This change affects how networks are managed and how customers are served. NVIDIA's latest State of AI in Telecommunications report shows 89% of respondents say open source models and software are important to their company's AI strategy. The reasons go beyond cost.

Five reasons operators cite

These models expand access to frontier-level intelligence at lower cost. Operators can reserve closed models for workloads where they drive the most value. Independent benchmarks such as the Artificial Analysis Intelligence Index v4.3.2 show leading open models becoming more competitive across demanding reasoning and coding tasks. Scientific and agentic workloads also see gains.

Customization is another driver. Open weights let operators fine-tune models for their own operations using network data. Customer data and industry data also feed the process. Training recipes are available as well. Trustworthy AI gets a boost because telcos gain greater visibility into and control over model artifacts and behavior. Models can be evaluated and adapted. They can also be governed in alignment with regulations and business policies.

Flexible, secure deployment rounds out the list. Teams can size and optimize open models to run across public clouds. Private infrastructure and edge environments are also options. The approach unlocks the opportunity for telcos to deliver locally adapted AI services to enterprise and government customers by hosting and fine-tuning open models.

Open foundations tuned for telecom

The NVIDIA Nemotron family of open models provides frontier-level reasoning performance optimized for agentic workflows. It also includes speech capabilities for voice applications. All come with open weights and training data. Recipes are included as well.

SoftBank Corp. illustrates how an operator can use open models as a basis for developing telecom-specific AI capabilities. "Open models allow SoftBank Corp. to build on the rapid progress of global foundation models while applying the network knowledge and operational expertise we have accumulated over many years," said Rajeev Koodli, principal fellow of SoftBank Corp. and senior vice president of SB Telecom America. The company uses open foundations extensively, including NVIDIA Nemotron models and others, in developing its SoftBank Large Telecom Model. This model is continuously advanced for telecom-specific use cases such as network operations and design. Overall management is another focus area.

NVIDIA announced the 30-billion-parameter Nemotron 3 Large Telco Model (LTM). It was fine-tuned by AdaptKey on open source telecom datasets to deliver accuracy gains for telecom-specific tasks. This gives operators an open baseline that can understand telecom industry terminology and reason effectively through telecom operations workflows such as network configuration and customer incident triage. To help operators customize the Nemotron 3 LTM and other open models with their own operational data, NVIDIA released the full recipe that walks through the end-to-end fine-tuning pipeline using NVIDIA NeMo open libraries.

From models to production workflows

Open models are a critical building block. But it takes more than models to bring autonomous telecom operations safely into production. For AT&T, the value of model choice lies in making that flexibility operational in alignment with its business priorities. "At AT&T, we believe the future of AI is not about choosing a single model; it's about intelligently matching every workload to the right combination of performance and cost," said Andy Markus, chief data/AI officer of AT&T. "Control is the third factor. Open models are essential to that approach, and with NVIDIA, we can bring that model-choice strategy into production with the scale and reliability our business requires. Governance is the final piece."

Operators need pipelines that prepare and protect their data for model fine-tuning by anonymizing sensitive records and generating privacy-preserving synthetic datasets. A platform that can turn open models into governed, autonomous agentic workflows is also required. NVIDIA provides that end-to-end platform, powered by NVIDIA AI Enterprise software and NVIDIA Agent Toolkit. The stack spans data pipelines and open models. Agent orchestration and secure runtimes form the core. Simulation completes the picture. Together with a strong partner ecosystem building at every layer of this platform, telecom operators have a clear path to translate the benefits of open models into production-ready AI workflows.

Local AI innovation on operator infrastructure

As telecom operators build AI infrastructure aligned with national AI strategies, open models give them a foundation to deliver AI services tailored to local languages and industries. Regulations and data governance requirements add further constraints. Operators across countries can fine-tune open models for local-language and industry-specific needs. They can also host open models on their trusted platforms that customers can consume directly or build on with their own data and applications.

For Indosat Ooredoo Hutchison, one of Indonesia's largest telecom operators, that means building AI that reflects the country's own language and culture through its Sahabat-AI family of open source models. "For countries like Indonesia, the value of open models goes beyond access to powerful AI. It is about adapting that intelligence to our own language, culture, data and real-world needs," said Chirag Sukhadia, chief data/AI officer of Indosat Ooredoo Hutchison. "Sahabat-AI puts this into practice, using open models as a foundation to build AI that understands Indonesia and can be developed for local applications. This allows us not only to adopt AI, but to build capabilities around it and enable more Indonesians to create with AI on their own terms."

The shift is underway. How many operators have already moved workloads into production on open foundations is not stated in the announcement.