Sovereign AI Agents: Local to Global AI Factories
European AI developers, take note: NVIDIA’s latest breakthroughs, showcased at VivaTech, are set to revolutionize the creation of secure AI agents. This critical advancement addresses key challenges, including compute limits and data privacy, crucial as the EU pours over $200 billion into AI. The new NVIDIA Enterprise AI Factory design provides a turnkey solution with Blackwell-accelerated infrastructure and a cutting-edge software stack. Utilize NIM capabilities for rapid inference across diverse open large language model (LLM) architectures; plus,discover AI Blueprints guiding creation,and onboarding of AI agents. This collaborative approach, supported by developer examples, ensures safety and continuous advancement.News Directory 3 provides even more insights into these exciting developments. Discover what’s next as enterprises scale AI factories globally, from financial services to healthcare?
NVIDIA AI Factories Boost AI Agent Progress in Europe
European Union AI developers are poised to benefit from new NVIDIA technologies designed to streamline the creation of secure, optimized sovereign AI agents. These tools address critical constraints such as limited compute availability, data-privacy needs, and safety priorities, as the EU invests over $200 billion in AI.
Announced at the NVIDIA GTC Paris keynote at VivaTech, the suite aims to simplify AI development and deployment.Enterprises can now build scalable AI factories on-premises or in the cloud to accelerate the creation of sovereign AI agents.
The expanded NVIDIA Enterprise AI Factory validated design offers a turnkey solution, pairing NVIDIA Blackwell-accelerated infrastructure with a next-generation software stack. A key component is a new NIM capability that enables rapid inference for various open large language model (LLM) architectures, supporting over 100,000 public, private, and domain-specialized model variants hosted on Hugging Face.
NVIDIA AI Blueprints and developer examples guide developers in simplifying the creation and onboarding of AI agents while ensuring safety, privacy, and continuous improvement. These tools include the Ai-q and data flywheel NVIDIA Blueprints, plus a blueprint for AI safety using NVIDIA NeMo.
Across Europe, major enterprises are already constructing NVIDIA-accelerated AI factories. BNP Paribas and Finance Computer Science are scaling AI factories to run financial services AI agents. L’Oreal-backed startup Noli.com is collaborating with Accenture, leveraging its AI Refinery for its AI Beauty Matchmaker. IQVIA is developing AI agents to enhance healthcare services.
BT group is optimizing customer service and addressing network anomalies in the telecom sector, while telenor is utilizing its AI factory to run NVIDIA AI Blueprints for autonomous network configuration.
“NIM makes it easy to deploy a broad range of LLMs from Hugging Face on NVIDIA GPUs,” said Jeff Boudier, vice president of product at Hugging Face. “With support for over 100,000 public and private LLMs hosted on the Hugging Face Hub, NIM makes the performance and diversity of open models available to enterprise AI agents.”
NIM microservices now support a vast collection of LLMs on Hugging Face, automatically optimizing models with inference engines like NVIDIA TensorRT-LLM, SGLang, or vLLM. This allows users to rapidly deploy their preferred LLMs for high-performance AI inference on any NVIDIA-accelerated infrastructure.
To further assist developers, NVIDIA offers blueprints and developer examples for building and onboarding secure AI models and agents. the Ai-q NVIDIA Blueprint guides the development of agentic systems for fast multimodal data extraction and information retrieval, incorporating the NVIDIA NeMo Agent toolkit.
The NVIDIA AI Blueprint for building data flywheels enables enterprises to continuously improve their AI agents by turning inference data into new training and evaluation datasets. The Agentic AI Safety blueprint offers a framework for evaluating and enhancing model safety across content, security, and privacy dimensions.
What’s next
Global enterprises,including ActiveFence,Amdocs,Cisco,Cloudera,CrowdStrike,IBM,IQVIA,SAP,ServiceNow,and Trend Micro,are adopting NVIDIA NIM microservices and blueprints to accelerate AI workflows across various sectors. Amdocs, for example, plans to integrate more NIM microservices to support AI agents for content creation, translation, network automation, and customer service.