Wind River Announces Strategic Collaboration with AMD to Unify Open RAN and AI-RAN on Shared Infrastructure
The industry's first commercially available platform to run Open RAN functions and AI workloads side by side on shared hardware — eliminating the duplicated infrastructure that has held back AI-RAN adoption at scale.
5 min read
Wind River, an Aptiv company and global leader in software for the intelligent edge, has announced a strategic collaboration with AMD to deliver what the companies describe as the industry's first commercially available platform capable of unifying open radio access network (Open RAN) functions and AI-powered radio access network (AI-RAN) workloads on shared hardware. The jointly engineered solution directly addresses one of the most persistent cost and complexity barriers facing telecommunications operators today: the requirement to maintain entirely separate infrastructure stacks for RAN and AI workloads — a duplication that can effectively double capital expenditure and significantly increase integration complexity.
The Problem With Today's AI-RAN Architecture
As telco operators have moved from Open RAN trials to commercial deployments, the aspiration to layer real-time AI capabilities onto their networks has collided with a structural constraint: traditional architectures treat RAN functions and AI inference as categorically different workloads requiring separate compute infrastructure. Traffic prediction, anomaly detection, energy optimisation, and network intelligence functions require a computing environment that is simultaneously deterministic enough for RAN latency requirements and flexible enough for AI inference — a combination that existing purpose-built systems were not designed to deliver on shared hardware.
By combining AMD EPYC™ CPUs with the Wind River Cloud Platform, the two companies have produced a production-ready solution that hosts virtualised RAN functions and AI inference workloads side by side on the same distributed platform — removing the need for the duplicated hardware that has constrained AI-RAN adoption to date.
"As operators move from Open RAN trials to commercial deployments, AI is becoming a critical capability, not a future add-on. Together with AMD, we are enabling customers to seamlessly add AI to their networks without doubling infrastructure, providing the intelligence they need without the complexity penalty."
— Javed Khan, Executive Vice President & President, Intelligent Systems, Aptiv
What the Unified Platform Enables
The jointly engineered platform gives operators four specific operational advantages over the separate-infrastructure approach:
- Reduced infrastructure costs and operational complexity — by consolidating Open RAN and AI-RAN workloads onto shared hardware rather than maintaining parallel systems, operators eliminate a significant and recurring source of capital and operational duplication
- Real-time AI capabilities at the network edge — including traffic prediction, anomaly detection, energy optimisation, and network intelligence running alongside virtualised RAN functions, without the latency penalties introduced by routing AI workloads to separate infrastructure
- Efficient scaling across thousands of distributed sites — with automated lifecycle management, resiliency, and fault tolerance built into the platform architecture, enabling large-scale deployments without proportional growth in operational overhead
- Architectural flexibility to evolve without hardware replacement — operators can introduce more advanced AI capabilities as requirements grow, with the platform's software-defined design preserving investment in existing infrastructure
"Telco-grade AI solutions require a flexible and high-performance compute foundation that can support both real-time RAN workloads and AI inference. Our world-class AMD EPYC CPUs deliver the performance and scalability to power AI-driven RAN architectures. Our collaboration with Aptiv and Wind River provides their customers a solution built on proven technologies so operators can deploy AI to support their business operations and requirements."
— Philip Guido, Chief Commercial Officer, AMD
Joint Engineering, Roadmap, and Live Demos at MWC Barcelona
The partnership extends beyond product integration into joint engineering focused on AI-RAN software stack optimisation and hardware-software co-optimisation, a shared development roadmap, and active proof-of-concept deployments with operator partners. Live demonstrations and full technical architecture overviews of the unified platform are available at MWC Barcelona 2026 at the Wind River booth in Hall 2, Stand 2F25.
Why This Matters for the 5G and Beyond Era
The Wind River and AMD collaboration arrives at an inflection point for the global telecommunications industry. Operators have invested heavily in Open RAN to break free from proprietary hardware lock-in — but AI has remained a parallel, disconnected investment that compounds rather than simplifies network complexity. By making AI a native, co-resident capability on the same distributed platform that runs RAN, this solution removes the most significant structural barrier to AI-RAN at commercial scale. For operators weighing the economics of next-generation network intelligence, the ability to add AI without adding infrastructure fundamentally changes the return on investment calculation — and accelerates the path from pilot to production deployment. More information is available at windriver.com.
Key Takeaways
- Wind River and AMD have launched the industry's first commercially available platform to unify Open RAN functions and AI-RAN workloads on shared hardware — eliminating the duplicated infrastructure that has blocked AI-RAN at scale
- The solution combines AMD EPYC™ CPUs with Wind River Cloud Platform to run virtualised RAN and AI inference workloads side by side on the same distributed platform
- Key operator benefits include infrastructure cost reduction, real-time AI at the network edge (traffic prediction, anomaly detection, energy optimisation), scalable lifecycle management across thousands of distributed sites, and architectural flexibility without hardware replacement
- The partnership includes joint engineering on AI-RAN software stack optimisation, a shared development roadmap, and active proof-of-concept operator deployments
- Live demonstrations and technical architecture overviews are available at MWC Barcelona 2026 — Wind River booth, Hall 2, Stand 2F25