Blue Yonder & NVIDIA build AI factory for supply chains
Tue, 19th May 2026 (Today)
Blue Yonder has developed a Model Training Factory with NVIDIA to create AI agents for supply chain operations.
The system uses NVIDIA Nemotron models and NVIDIA NeMo tools to fine-tune and test specialised models for warehousing, planning, transport, merchandising and network operations.
Blue Yonder describes the training system as a repeatable way to produce models that can handle specific supply chain workflows and support human operators. Models are graded before deployment and refined over time against set evaluation standards.
Warehouse agents
The first models are aimed at warehouse management workflows, including allocation shorts, inventory exceptions, due-time urgency, and inventory across yard and receiving trailers. These are areas where operators often need to make rapid decisions as conditions change during a shift.
The approach reflects a broader shift in business software from general AI assistants to groups of narrower agents that can reason through tasks, use tools and act within defined workflows. In supply chains, those decisions often need to be made in real time across warehouses, transport lanes and stores.
The Model Training Factory uses a hybrid setup, combining larger general models for broader reasoning with custom-trained supply chain models for more precise, lower-cost workflows. Nemotron's range of model sizes allows different models to be matched to different tasks, from frequent warehouse decisions to more complex planning work.
The models are trained on synthetic data rather than customer data. Blue Yonder is also using NVIDIA AI Enterprise for infrastructure management and software support in the training environment.
Warehouse focus
Warehouses are the first proving ground for the system because operations can change quickly. Late arrivals, equipment failures and shifting priorities can force managers to rework allocations and task schedules throughout the day.
Blue Yonder said a specialist agent can assess far more trade-offs in a short period than a human operator typically can, potentially making continuous decision support practical across warehouse networks. Later models are expected to be applied across the wider Blue Yonder software portfolio.
The training factory is built around Blue Yonder's existing operational data, workflow logic and subject-matter expertise gathered over decades in supply chain software. The company said this loop of workflows, telemetry, expert input and retraining creates a reusable method for producing additional models across different parts of the supply chain.
Duncan Angove, Chief Executive Officer at Blue Yonder, said the company's research had shown the limits of large general-purpose models in day-to-day supply chain decision-making.
"Supply chain has always been an AI domain. Our research into how agentic models perform on real warehouse and planning decisioning is exactly why we know where frontier models hit a wall," said Duncan Angove, Chief Executive Officer, Blue Yonder.
"Working with NVIDIA, we're building owned intelligence, not rented intelligence-supply chain models trained on the workflows, telemetry, and decision logic that actually run a warehouse or a planning system. This isn't a one-off fine-tuned model. It's a factory, and it produces purpose-built agents at the speed, precision and cost the autonomous supply chain demands," said Angove.
Domain models
NVIDIA presented the project as part of the next stage of AI adoption in large companies, with domain-trained agents embedded into operational systems rather than used only as standalone assistants.
"The next phase of enterprise AI for supply chains requires specialised, affordable and accurate domain-trained agents that can operate within the workflows that run a business," said Azita Martin, Vice President and General Manager, Retail and CPG, NVIDIA.
"Blue Yonder is leveraging NVIDIA Nemotron, the NVIDIA NeMo Agent Toolkit and NVIDIA AI Enterprise to build a Model Training Factory that fine-tunes models with proprietary supply chain data, enabling them to build agentic AI systems for some of the world's largest and most complex supply chains," said Martin.
The first models are expected to enter customer production through Blue Yonder's Cognitive Solutions offering later this year. The company serves more than 3,000 retailers, manufacturers and logistics providers.