NTT, SK Group, and Chunghwa Telecom have committed $500 million to a fund focused on AI infrastructure built around the IOWN photonics architecture.
Called the IOWN AI Fund and managed by the newly established IOWN Global Capital, the vehicle invests in companies building on the Innovative Optical and Wireless Network (IOWN) framework. The fund aims to accelerate commercial deployment of an optical-first approach to data transport—prioritising photonics over electronics—before cloud hyperscalers set the dominant standard.
Deconstructing the IOWN architecture
IOWN rests on three technical pillars, each designed to remedy a specific shortcoming of current networks.
The first pillar is the All-Photonics Network (APN). Today’s networks convert optical signals to electronic form at each node and then back to optical for onward transmission. Those conversions are fast but expensive: they consume power and add latency. APN keeps data in the optical domain along the entire path, from core to terminal device, transmitting light rather than electrons. This end-to-end optical approach reduces latency and energy consumption per bit—crucial for AI inference workloads that require sub-millisecond responses.
The second pillar, Digital Twin Computing (DTC), targets continuous, bidirectional data exchange between physical assets and their virtual replicas. Production-grade digital twins—used to simulate factory assembly lines or model distribution networks—need sustained, low-latency throughput that current networks struggle to deliver. DTC, when run over the APN layer, is engineered to support those high-volume, real-time data flows.
The third component, the Cognitive Foundation, operates above APN and DTC as an autonomous orchestration layer. It dynamically allocates network slices, compute, and storage based on real-time application demand. Designed to be self-healing and self-tuning, the Cognitive Foundation eliminates the need for manual intervention and turns the network into a programmable service with predictable performance—an important shift for enterprises deploying AI at the edge.
Operational calculus of photonics
The consortium’s $500 million investment acknowledges mounting operational pressures on the digital economy. Rapid growth in AI models and rising data traffic are straining power grids and legacy network infrastructure. Data centres packed with electronic switches and processors are becoming major energy and performance bottlenecks. IOWN’s photonics-electronics convergence tackles that challenge by shifting transmission to optical components, which move data with far lower energy per bit than electronic equivalents.
NTT holds the core architectural intellectual property and formed the IOWN Global Forum to steward the framework’s standards. SK Group, with semiconductor supply-chain capabilities, and Chunghwa Telecom, Taiwan’s largest carrier, expand the initiative beyond a single operator’s proprietary project into a potential regional standard. These partners bring manufacturing capacity, carrier infrastructure, and procurement scale across major technology-producing markets. The Development Bank of Japan’s involvement underscores how the initiative blends venture capital with strategic industrial policy, positioning IOWN as an infrastructure priority.
Enterprise implications beyond bandwidth
Use cases such as telesurgery, remote industrial robotics, and autonomous-vehicle control require guaranteed round-trip latencies measured in single-digit milliseconds. By removing optical-to-electronic conversion, APN makes those latency guarantees technically feasible. DTC enables enterprises to move from predictive maintenance to prescriptive optimisation: a manufacturer could run a live digital replica of a factory floor, ingesting real-time sensor data from hundreds of IoT nodes to test production changes virtually before implementing them physically. That capability reduces unplanned downtime and speeds iteration, but it depends on a network that can sustain continuous, high-volume data streams without the delays of today’s cloud-routed deployments.
The Cognitive Foundation also changes how enterprises procure network services. Currently, organisations seeking guaranteed performance must choose expensive dedicated leased lines or accept best-effort public internet links. The CF model adds a third option: programmable, autonomous network slices with defined service characteristics, provisioned dynamically and without human intervention. This converts connectivity from a commodity utility into a deterministic element of the enterprise IT stack, enabling reliable edge AI inference for factory sensors, security systems, and autonomous equipment.
Telco-led counterpoint to hyperscaler dominance
Telecom operators have watched cloud providers capture the higher-value layers of the internet economy. Under the cloud era model, telcos remained responsible for physical transport—absorbing capital costs and providing capacity—while hyperscalers captured service-layer value. The shift toward AI has deepened this imbalance, since model training and inference now largely occur inside the data centres of a few major technology companies rather than within telecom networks.
The IOWN AI Fund represents a strategic response. By integrating compute and networking at the architectural level—designing the infrastructure to host compute as an inherent capability rather than bolting it on—telcos and their partners aim to create a value proposition that hyperscalers cannot simply buy with bandwidth. Distributed AI applications that require real-time local processing rely on network properties centralised cloud infrastructure cannot easily provide. The fund’s mission is to build the vendor and startup ecosystem that can deliver those properties commercially.
Whether this strategy succeeds depends on adoption timing and the consortium’s ability to secure developer and enterprise commitment before cloud-native alternatives evolve. The $500 million commitment is significant: it exceeds the annual R&D budgets of many telecom operators outside the global top tier, reflecting both the cost and ambition of attempting to shift infrastructure paradigms.
IoT News is produced by TechForge Media and highlights enterprise technology trends, events, and webinars across the sector.