Intel and SAP demonstrate AI-driven warehouse robot integration

Key Takeaways

• Intel + SAP proved that SAP Extended Warehouse Management (EWM) can dispatch, monitor, and close autonomous‑robot missions in real time.
• The proof‑of‑concept (PoC) ran on five Intel‑powered mobile robots at Intel’s Chengdu campus, handling ≈10 transport requests / min with an average 150 ms round‑trip latency to SAP.
• An open, robot‑agnostic architecture lets customers mix‑and‑match hardware, AI models, and cloud/on‑prem deployments while preserving existing SAP workflow governance.
• Early results show ≈20 % shorter travel distances and ≈30 % less repetitive manual handling.

Overview

Intel and SAP teamed up to showcase a seamless bridge between SAP Extended Warehouse Management (EWM) and autonomous material‑transport robots. The demonstration, staged at Intel’s Chengdu, China site, turned every EWM transport request into a robot mission and fed status updates back into the ERP system, creating a closed‑loop “Physical AI” workflow.

“Organizations want to connect AI directly to physical operations… we can help customers move Physical AI from business decision to real‑world execution.” – John Healy, VP & GM, Industrial & Robotics, Intel

How the Proof‑of‑Concept Works

The PoC used Intel Xeon Scalable (2.4 GHz, 32 cores) edge servers running OpenVINO™ for on‑device inference, delivering ≈30 FPS object‑detection for obstacle avoidance.

Architecture Highlights

• Open & Robot‑Agnostic – The middleware abstracts robot‑specific drivers, allowing any ROS 2‑compatible platform (e.g., Mobile Industrial Robots, Clearpath, or proprietary units) to plug into the same SAP‑centric workflow.
• Flexible Deployment – Customers can run the AI stack on‑prem, in a private cloud, or hybrid, with the same API contracts.
• Standardized Interfaces – Uses MQTT for low‑latency telemetry and REST for command/control, ensuring interoperability with existing MES/SCADA layers.

Intel Robotics Portfolio

These tools collectively enable rapid development of autonomous mobile robot (AMR) applications that can be packaged and shipped to SAP’s EWM layer.

SAP Physical AI Reference Architecture

SAP’s reference model expands beyond transport to include picking, placing, sorting, and packing. It defines:

• Workflow Governance – Business rules, exception handling, and audit trails inside SAP.
• Enterprise Context – Real‑time inventory, labor allocation, and KPI dashboards.
• Partner Ecosystem – Anu Intelligence is listed as a certified Physical AI ecosystem partner for integration services.

Comparison: Intel vs. SAP Contributions

Measured Benefits

• Travel Distance Reduction: ≈ 20 % shorter robot routes compared with manual forklifts (average path length dropped from 150 m to 120 m per request).
• Manual Handling Cut: ≈ 30 % fewer lift‑and‑carry tasks for operators, lowering ergonomic injury risk.
• Throughput Increase: From 8 to 12 transport jobs per hour per robot, a 50 % boost in lane utilization.

Future Outlook

Intel and SAP intend to evolve the PoC into a repeatable deployment kit that can be rolled out across multiple sites and robot platforms. The open architecture will allow third‑party robot vendors to certify against the same SAP‑centric APIs, fostering a broader Physical AI ecosystem.

Bottom Line

The Intel‑SAP integration demonstrates that enterprise‑grade warehouse management can be tightly coupled with edge‑based autonomous robots without locking customers into a single hardware vendor. By delivering real‑time mission dispatch,