NVIDIA has announced the general availability of its NeMo microservices, an integrated developer platform designed to streamline the creation and deployment of agentic Artificial Intelligence systems within enterprise IT environments. These microservices allow organizations to build and manage AI-powered digital teammates that leverage user interactions, inference data, and business intelligence to continuously enhance model performance through orchestrated data flywheels—systems that curate, retrain, and evaluate models with minimal human input.
Key components of the NeMo suite, including NeMo Customizer, NeMo Evaluator, NeMo Guardrails, NeMo Retriever, and NeMo Curator, provide tools for fine-tuning large language models, benchmarking against both custom and industry standards, and enforcing robust safety and compliance protocols. For example, enterprises can accelerate large language model training throughput by up to 1.8x and improve compliance protection by 1.4x. NeMo microservices are deployable on NVIDIA AI Enterprise and operate seamlessly on any accelerated computing infrastructure, offering enterprise-grade security and support either on-premises or in the cloud.
Industry leaders such as AT&T, BlackRock, Cisco’s Outshift team, and Nasdaq are leveraging NeMo microservices to build advanced AI agents that deliver measurable gains: AT&T used the platform to increase agent accuracy by 40% for core business functions such as personalized service and fraud prevention, while Nasdaq improved its platform’s response time and accuracy by up to 30%. The platform supports a broad array of foundation models, including Meta’s Llama, Microsoft Phi, Google Gemma, Mistral, and the Llama Nemotron Ultra, enabling versatile development across diverse enterprise needs.
NeMo microservices are now integrated by a growing ecosystem of technology partners and solution providers, from major storage and infrastructure companies like Dell Technologies, HPE, IBM, NetApp, and Pure Storage, to software platforms and consulting leaders, including Accenture, Deloitte, and EY. Developers can access NeMo microservices via the NVIDIA NGC catalog, with full API support for popular frameworks and guarantee of stability, proactive security updates, and support through NVIDIA AI Enterprise’s extended-life software branches.
