The industry is shifting toward AI-native networks, but in most live deployments today, RAN and AI workloads still run on separate hardware stacks – different servers, different software, no unified orchestration. That fragmentation drives up CapEx and OpEx, slows innovation at the edge, and makes it hard to launch AI-based services at scale.
SynaXG has built a commercial-ready AI-RAN platform that runs 5G FR1, 5G FR2 and AI workloads together on the same cloud-native, GPU-accelerated infrastructure, with real-time orchestration of resources between them.
Today, a typical deployment is more siloed architecture which looks like this:
- A 5G FR1 stack for macro and indoor coverage
- A separate 5G FR2 / mmWave stack for hotspots and FWA
- A separate AI / GPU stack for large models and other AI apps
Each stack has its own hardware, lifecycle and operations. This model does NOT scale:
- Low hardware utilization and duplicated investment
- Complex operations across multiple platforms
- Slow roll-out of AI and edge services because connectivity and AI are not designed to work together
Our Answer: One Unified AI-RAN Platform
SynaXG’s platform is designed from day one to host RAN and AI together:
- 5G FR1 and 5G FR2 on the same platform
- Layer 1/2/3, vCU/vDU
- Support for both sub-6 GHz and mmWave in a unified software stack
- AI workloads on the same hardware
- AI-for-RAN: energy saving, interference management, link adaptation and other RAN optimization agents
- AI-on-RAN: Large models, Robotic Inspection, Vision AI, Drone Orchestration and other edge applications
- Naturally enables AI-and-RAN scenarios where AI and RAN cooperate in real time
- Cloud-native and open API
- Built on Kubernetes (e.g. Red Hat OpenShift)
- Open interfaces so operators, hyperscalers and enterprises can go onboard with their own AI applications and services
- Instead of three parallel stacks, one platform now serves connectivity, capacity and AI workloads at the same time.
Real-Time Orchestration
The key innovation is not just consolidation. It’s real-time orchestration between AI and RAN:
- Compute resources (CPU/GPU) are allocated dynamically between FR1, FR2 and AI workloads based on traffic, policy and SLAs.
- RAN performance is ensured (throughput, latency, reliability) while AI workloads receive guaranteed capacity for their own SLAs.
- The platform supports fine-grained partitioning and scheduling, so operators can:
- Consolidate multiple cells and AI applications on the same node
- Shift capacity between AI and RAN as demand patterns change over time
This turns the platform into a programmable infrastructure layer where business logic (tenant, slice, service type, priority) directly influences how resources are shared between 5G and AI.
Proof Points and Ecosystem Traction
We are validating this architecture with:
- Lab deployments where 5G FR1, 5G FR2 and AI workloads run concurrently on GPU servers while maintaining strong RAN KPIs, with commercial-level traffic patterns and use cases running stably.
- Industry collaborations with operators, cloud partners and system integrators through initiatives such as TM Forum Catalysts, where SynaXG serves as the AI-RAN platform provider.
These projects give us practical data points on performance and new AI use cases, and they provide a clear path from proof-of-concept to commercial rollout.
Commercial Progress
Beyond SynaXG’s continuous breakthroughs in core technology and product innovation, it has also achieved significant commercial progress.
In September 2025, SynaXG signed global Partnership Agreement with Tech Mahindra, strengthening the companies’ joint commitment to accelerating AI-RAN and Private 5G (P5G) adoption worldwide.
Building on SynaXG’s continuous breakthroughs in core technology and product innovation, the partnership combines SynaXG’s advanced AI-RAN and end-to-end connectivity solutions with Tech Mahindra’s global system-integration expertise and deep enterprise reach. Together, the companies will promote next-generation intelligent connectivity solutions for Industry 4.0 such as smart manufacturing, smart ports, smart campuses, and other Industry environments. Under the agreement, Tech Mahindra will serve as the lead systems integrator, enabling joint bids, coordinated delivery, and integrated lifecycle support for customers globally. This partnership marks a major milestone in accelerating large-scale commercial deployment of AI-driven P5G networks worldwide.
In November 2025, SynaXG joined TM Forum Innovate Asia 2025 as the AI-RAN platform partner in the Catalyst project “Project AURA (AI-RAN)”, working alongside StarHub, Etisalat UAE, CelcomDigi, OREX SAI and Red Hat.
The joint project demonstrates how operators can move from connectivity to AI-first, revenue-generating infrastructure at the network edge.
About SynaXG
SynaXG Technologies is a Singapore-based deep-tech company pioneering AI-RAN and next-generation RAN solutions. Founded by wireless veterans with decades of domain expertise, SynaXG delivers full-stack RAN software (L1/L2/L3), vDU/vCU, RU platforms, and end-to-end RAN systems optimized for heterogeneous compute architectures across GPU, CPU, FPGA, ASIC and more. The company works with global technology leaders and telecom operators to build the future of AI-native wireless networks.