The Intelligence Has Always
Been in the RAN.
Now You Can Use It.
Scroll to explore
The Problem We Solved
Vast Infrastructure — Intelligence Left on the Table
Every base station continuously generates intelligence across its full software stack — beam metrics, channel estimates, interference signatures, spatial measurements, Doppler data, IQ streams, and mobility events. This intelligence is rich, real-time, and — until now — inaccessible to the applications that should act on it.
RANIQ turns every base station into a programmable RF sensor node — unlocking intelligence already present in the network, without requiring new RF hardware or spectrum.
The industry has made meaningful progress in opening up the RAN. O-RAN interfaces, RIC-based applications, SMO analytics, and vendor telemetry hooks have all moved the needle. But these interfaces were designed for network management and optimization — they expose aggregated, post-processed outputs. Data that has already been filtered, delayed and stripped of the sub-millisecond temporal precision that real-time intelligence applications require.
The deeper challenge is integration depth. Accessing signal-level data — controlled IQ sample windows, slot-synchronized beam weights, live channel estimation matrices — requires intimate coupling with the RAN’s own signal processing pipeline. That cannot be achieved from a management layer. It has to be designed in from the start by the same team that built the RAN stack itself.
RANIQ was built to close that gap: a platform that integrates at the right depth, exposes the right data in a controlled and governed manner, and makes it consumable by applications without the complexity or integration overhead of external approaches.
"The RAN is not just a pipe. It is a sensor array, a signal processor, and an AI inference engine — when the platform is designed from the start to treat intelligence as a first-class output."
Platform Architecture
RANIQ: Native Intelligence at the Radio Edge
RANIQ is ODC’s Unified Edge Intelligence Platform — co-resident with ODC’s RAN software on the same DU/CU compute node. It is not middleware. It is not a management-layer analytics tool.
It is an intelligence fabric co-resident on the same compute node as the radio stack, with controlled access to platform resources and strict isolation from RAN real-time threads. Applications built on RANIQ run on that same node, giving them direct, low-latency access to RAN intelligence without any network round-trips.
RANIQ operates as an independent runtime on the DU/CU compute node — managing data access, resource isolation and DApp lifecycle. The RANIQ SDK abstracts all aspects of the interface to RANIQ — data access, resource management, lifecycle and runtime interaction — so DApp developers work entirely within the SDK without any direct exposure to the underlying platform internals.
Understanding where RANIQ sits in the stack matters. It resides between the RAN signal processing pipeline and the application layer — capturing intelligence at the point of generation and making it available through structured APIs before it is filtered or discarded upstream:
(CSI · IQ · Beam · Mobility) →
All three components are co-resident on the same DU/CU compute node while maintaining strict isolation from RAN real-time processing. DApps interact with RANIQ exclusively through the RANIQ SDK, which abstracts all platform internals.
This co-residency is the central design decision that differentiates RANIQ from external analytics or O-RAN-layer approaches. By integrating directly alongside the RAN L1/L2/L3 stack, RANIQ can access data at the resolution and latency that signal-level applications require — before it is aggregated, filtered, or discarded by higher-layer processing.
Data is exposed as controlled subsets: sample windows, feature tensors, and derived signal metrics — not unlimited raw streams. This keeps memory bandwidth consumption bounded and security exposure within operator-defined limits.
RANIQ vs Management-Layer Approaches
O-RAN xApps, RIC-based applications and vendor telemetry interfaces have made the RAN more programmable — and they remain the right tool for network management tasks. They operate at the KPI and counter layer: periodic reports, aggregated measurements, control-plane events.
RANIQ operates at a different layer with a different purpose: real-time signal intelligence for applications that need it at sub-slot resolution. The two are complementary, not competing.
RANIQ’s value is specifically in the class of applications — sensing, real-time ML inference, interference analytics and distributed ISR — where management-layer latency and data resolution are insufficient. It is not a replacement for O-RAN extensibility. It is an additional integration layer that makes a specific, high-value class of applications possible on existing RAN compute.
Pillar 01
Real-Time RAN Intelligence Fabric
A controlled, high-throughput data layer exposing beam and CSI insights — Channel State Information, the measure of radio path quality between antenna and device — resource maps, mobility events, timing measurements and derived IQ features. Published as event-driven or poll-based streams to subscribing applications.
Pillar 02
ML Inference Platform
Slot-aligned inference pipelines operating on real-time RF feature streams, co-resident on the same compute node as the RAN. Model registry with over-the-air hot-swap. Supports CPU and GPU inference paths.
Pillar 03
Signal Processing Toolkit
Advanced RF analytics as first-class APIs: Doppler estimation — detecting velocity from frequency shift of radio signals — interference classification, beamforming taps, FFT pipelines, and ODC’s proprietary RF algorithms. Domain expertise, pre-packaged.
Pillar 04
Cloud-Native DApp Runtime & SDK
A cloud-native DApp SDK with full lifecycle management (LCM), built-in observability, and auto-healing. DApp workloads execute in containerized environments — with isolated or shared/pooled resource allocation depending on the sensitivity and resource profile of the application. CPU-core pinning is supported where deterministic latency is required but is not mandated for all workloads.
What RanIQ Unlocks
Applications That Require Signal-Level Intelligence
Real-time access to RAN-native intelligence — across signal, beam, mobility, and resource dimensions — enables a class of applications that management-layer data alone cannot support. These are engineering problems that become tractable when the data is rich enough, fast enough, and structurally close enough to the radio.
RANIQ provides the platform foundation — the data access, processing primitives, runtime, and flexible DApp isolation architecture. Meaningful applications still require domain expertise in RF, signal processing, or ML. RANIQ removes the integration barrier; it does not replace the engineering work of building on top of it.
RF-Based Awareness
Using the RAN as a distributed sensor array to detect, localize, and classify objects and movements — no additional hardware, no extra spectrum. Applicable to airspace monitoring, perimeter security, and industrial safety.
Analytics & Intelligence
Spatial & Mobility AI
Real-time occupancy mapping, crowd flow analysis, and asset tracking — derived from RF-native mobility intelligence. Privacy-preserving by design: no user-identifiable data, operating under operator data governance policies.
Defense & Security
Distributed RF Sensing & ISR
RANIQ enables a distributed RF sensing fabric using commercial RAN infrastructure — wide-area RF awareness, emitter detection, RF fingerprinting and interference attribution — without dedicated radar hardware.
Applications include: drone & UAV detection · RF emitter geolocation · spectrum monitoring · battlefield RF awareness
Network AI
Self-Optimizing RAN
ML models trained on real signal-level data — not just KPI counters — driving predictive handover optimization, interference mitigation, and proactive load balancing. Inference timescales from sub-slot to minutes, depending on the application.
Industrial & Enterprise
Private RAN Intelligence
Safety monitoring, sub-meter positioning, process automation, and anomaly detection — running on the same compute already at the RAN edge. Both CPU-based analytics and GPU-accelerated ML inference are supported natively, with no additional edge servers required.
Spectrum Resilience
Anti-Jamming & Interference Mitigation
Real-time interference classification, jammer detection, and adaptive null-steering — powered by RANIQ’s access to controlled IQ sample windows and beam weight vectors. Spectrum resilience embedded into the RAN compute layer itself.
Custom Applications
Built on RANIQ SDK
Teams with RF, signal processing, or ML expertise can build domain-specific applications on RANIQ without engaging with RAN stack internals. ODC manages the radio integration; application teams own the domain logic.
RANIQ enables a developer ecosystem where RF analytics, ML models, and sensing applications can be deployed across operator networks as portable DApps — composable, observable, and manageable through the RANIQ Design Studio.
What these use cases share is a dependency on RAN intelligence at the right resolution for the application. RANIQ supports workloads across the full spectrum — from sub-slot signal processing and millisecond ML inference, through to multi-second analytics pipelines — on the same platform, under the same resource isolation architecture. Management-layer interfaces were not designed to provide data at this depth or latency. RANIQ is built specifically for this class of requirement.
Proof of Platform · ODC Innovation
ISAC: The Application That Proves the Foundation
ISAC — Integrated Sensing and Communication — is one of the most technically demanding application classes in wireless, combining active sensing with communication in an integrated system. As a 3GPP standardization area, ISAC is still maturing. But the platform requirements it places on the RAN are concrete — and RANIQ was designed to meet them.
Why ISAC Is the Right Platform Validation
An ISAC deployment on RANIQ uses 3GPP-standard waveforms already present in the network as sensing signals, routing intelligence through RANIQ’s signal processing toolkit, ML inference platform, and distributed data fabric to produce real-time sensing outputs — concurrently with mobile service, within the same radio slot structure, under controlled resource partitioning between sensing workloads and RAN L1 processing.
ISAC demands the deepest possible platform integration: controlled IQ sample access at sub-slot timing, deterministic compute scheduling enforced through CPU isolation and priority queuing, co-resident signal processing with bounded latency, and strict separation from RAN real-time threads across L1 and L2. These are precisely the engineering challenges RANIQ was designed to solve. A platform that meets ISAC’s requirements provides sufficient headroom for analytics, ML inference, and monitoring applications operating at less demanding timescales.
Every application built on RANIQ operates within a less demanding envelope than ISAC requires. Validating the platform architecture at the ISAC level provides engineering confidence across the full application tier above it.
"ISAC sets the ceiling on what the platform needs to handle. Every other application operates below that ceiling — with the same architecture, the same resource isolation, and the same engineering rigor underneath."
Operator Opportunity
Your Network. A Second Dimension of Value.
Mobile operators already run the world’s largest distributed RF sensor network — RANIQ turns it into an intelligence platform, enabling new service opportunities using existing infrastructure.
Operator monetization of edge and intelligence capabilities is not a new ambition. MEC platforms, network slicing, and edge marketplaces have explored adjacent ground with limited commercial success. The recurring challenge has been the gap between platform capability and enterprise demand. RANIQ’s positioning addresses this directly: rather than a general-purpose edge compute play, it focuses on intelligence derived from the RAN itself — something operators uniquely possess and that has identifiable demand in specific verticals today.
The clearest near-term opportunities are where the customer relationship and use case are already defined — private RAN deployments, defense and critical infrastructure contracts, and industrial operators with specific sensing or analytics requirements. These contexts bypass the demand-development challenge that has slowed broader edge monetization.
Opportunity 1
RF Environment Intelligence Services
Offer RF sensing outputs — spectrum monitoring, perimeter detection, airspace RF awareness — as subscription services to enterprise and government customers. Concrete service categories with identifiable buyers and existing budget lines, particularly in defense and critical infrastructure.
Opportunity 2
Mobility & Spatial Analytics Platforms
Provide anonymized, aggregated mobility intelligence — crowd flow, occupancy analytics, asset movement — to smart city integrators, transportation operators, and retail enterprises. Data governance and privacy frameworks are prerequisites; RANIQ’s architecture supports their implementation.
Opportunity 3
Industrial Safety & Spectrum Monitoring
For private RAN operators in industrial environments: RF-based safety zone monitoring, equipment proximity detection, and spectrum health services — running on existing RAN compute, with no additional sensor infrastructure required.
"Operators have a data asset that no cloud provider owns — real-time RF intelligence from a distributed network of radio nodes. RANIQ makes that asset accessible, structured, and consumable. What the market builds on top of it will define the next phase of network value."
What Makes RANIQ Possible
Built From the Inside Out
RANIQ is built by the same engineering team that built ODC’s RAN — and that integration depth is what makes the platform architecture possible. These are engineering decisions with specific rationale, not marketing claims.
-
✦
Co-resident with the RAN, not layered above it — RANIQ runs co-resident on the same DU/CU compute node as ODC's RAN L1/L2/L3 stack while maintaining strict isolation from real-time RAN processing. This placement enables signal-level data access before aggregation — a fundamentally different integration point than O-RAN management-layer interfaces, which operate on post-processed outputs.
-
✦
Flexible DApp isolation and resource management — DApp workloads support both isolated execution (with dedicated CPU core pinning, NUMA-aware memory allocation, and accelerator scheduling controls) and shared/pooled resource architectures for better utilization across less latency-sensitive workloads. RAN real-time threads (L1 and L2) hold strict scheduling priority regardless of the DApp resource model. Data can be exchanged with DApps on the same compute node or across compute nodes via a fast-path transport layer, with marginal latency increase for cross-node deployments.
-
✦
Slot-aligned timing architecture — RANIQ captures data at slot boundaries — the fundamental timing unit of the radio frame — with buffering layers that absorb processing jitter while maintaining temporal alignment. Deterministic execution scheduling ensures sensing and inference workloads do not interfere with L1 timing.
-
✦
Controlled data exposure, not unlimited streaming — RANIQ exposes intelligence as controlled subsets: sample windows, derived feature tensors, and signal metrics. Raw IQ access is scoped and governed. This keeps memory bandwidth consumption bounded and reduces security exposure to operator-defined limits.
-
✦
RANIQ Design Studio for managed solution composition — RANIQ includes a Design Studio that enables operators and integrators to build managed solutions as a composition of DApps — configuring data flows, resource policies, and application topology without engaging with low-level runtime details.
-
✦
Privacy and data governance by design — RANIQ operates on RF and signal data, not user-identifiable content. The SDK supports anonymization pipelines, developer permission tiers, and operator-configurable data access policies. Applications are responsible for their own regulatory compliance; RANIQ provides the governance infrastructure to support it.
-
✦
ODC's RF signal processing as platform APIs — Years of ODC R&D in Doppler estimation, interference classification, and beamforming are packaged as high-level API calls. Application teams work with structured signal intelligence, not raw radio mathematics — lowering the domain barrier for teams with ML or systems backgrounds.
-
✦
Runs on existing DU/CU compute — no new hardware — RANIQ and its application runtime deploy as software on the same compute node already running ODC’s RAN stack, developed on the NVIDIA AI Aerial platform — no dedicated edge servers, no additional hardware. CPU-only deployments equally supported.
-
✦
Requires ODC's RAN as foundation — RANIQ is available to operators and enterprises running ODC's RAN software. This is a deliberate integration choice, not an arbitrary dependency. For those evaluating RAN vendors, RANIQ represents a platform dimension that is worth factoring into the selection decision.
RANIQ reflects architectural decisions made from the beginning of ODC’s RAN development — treating intelligence as a first-class output alongside connectivity — not as a retrofit, but as the result of building both the RAN and the intelligence platform as a unified system from day one.
The RAN Has Always Been
More Than a Radio.
RANIQ is available today. Whether you are an operator evaluating new revenue surfaces, an enterprise building on private RAN, a defense integrator exploring distributed RF sensing, or a developer with a use case that needs signal-level RAN intelligence — we want to talk.