(Information sent by the signing company)
VIENNA, October 16, 2026 /PRNewswire/ -- In the framework of the Network X 2026 event, the World Broadband Association (WBBA) held the Broadband Development Congress (BDC) in Vienna. During this meeting, Bob Chen, president of Huawei's optical solutions product line, delivered a keynote titled "Reconfiguring Network Value for New Growth in the AI Era." In his speech, he emphasized the urgency of updating network building approaches in the AI era based on three "R" pillars —rebuilding network architecture, reconfiguring the business model, and redefining operation and maintenance (O&M)— with the aim of rethinking network value and opening a new stage of growth.
Bob Chen recalled that, if we look back over the last thirty years, each wave of innovation in services —from voice telephony and Internet access to video— has always required a solid network as a foundation. Currently, AI is advancing at great speed: large AI models already handle nearly 10 trillion parameters and, just in the last year, token consumption has skyrocketed by almost 50. In this new context, connected objects, usage frequency, and connection quality have changed radically, forcing the sector to reconsider how to reconfigure network value.
The executive insisted that the core of this reconfiguration involves modifying the deployment philosophy: the planning and construction of networks must revolve around the needs of AI services and the location of AI data centers (AIDC), also incorporating AI technology itself to reinforce O&M capabilities. This vision is articulated in three "R" pillars: reconstructing the network architecture, migrating from structures based on classic administrative divisions to designs centered around AIDCs; reconfiguring the business model, so that, in the business realm, it shifts from solely marketing bandwidth to offering "bandwidth + latency + reliability" along with integrated computing and network services, and that, in the residential market, it evolves from mere connectivity to whole-home networks and, subsequently, to complete smart home experiences; and, finally, redefining operation and maintenance, advancing from semi-automated schemes to an intelligence that covers the entire lifecycle of the network.
Reconstruction of the network architecture
Traditionally, networks were designed based on administrative boundaries, with divisions between national trunk networks, provincial trunks, and metropolitan networks. With the rise of AI, the number of AI data centers (AIDC) has multiplied almost tenfold and the volume of traffic no longer depends so much on the population of each area, but on the installed computing capacity.
An illustrative example is Zhongwei, in China: a few years ago, its traffic accounted for barely around 0.1% of the national total. Currently, Zhongwei has become one of the eight national computing centers in the country and its intelligent computing traffic exceeds 10% of the total bandwidth of China, transforming it into a key node within the AIDC interconnection network. Similar situations are being recorded in multiple regions of the planet.
In light of this change, the network architecture must adapt and shift from a design linked to administrative regions to one that references the AIDC. Networks must be structured in hierarchical levels articulated around central data centers (Core DCs), regional (Regional DCs), and edge (Edge DCs). The one-hop mesh connections between these centers allow for the establishment of deterministic latencies —of 1, 5, and 20 ms— and accelerate the time to market for new services.
Reconfiguration of the business model
In the enterprise segment, the push for AI is transforming production models. For example, a Chinese industrial design company based on AI uses AI rendering to drastically shorten production cycles, which requires latencies of less than 2 milliseconds. To capitalize on these opportunities, operators can evolve their private lines, leaving behind a proposal based solely on bandwidth and moving to a multidimensional offering: "Bandwidth + Latency + Reliability." Through technologies like OXC, ASON, and quantum encryption, networks are capable of providing deterministic latency, high reliability, and a maximum level of security, significantly increasing the value of private lines. At the same time, by adding computing platforms, operators can make the leap from selling only connectivity to a model of "Network + Computing + Model/Application," offering convergent network and computing services.
In the realm of residential broadband, operators have the option to advance from simple Internet access provision to a strategy of "One fiber, one network, one home." A network implies ensuring quality Wi-Fi connectivity in all rooms through FTTR solutions or mesh Wi-Fi networks. A home means providing users with a broad ecosystem of smart applications, supported by home terminals with AI and AI agents. These agents are changing the way of interacting in the home environment, allowing real-time management of the residential network through simple voice commands.
Redefining network operation and maintenance
In the midst of the AI era, an integral intelligence is pursued that encompasses the phases of planning, deployment, maintenance, optimization, and operation, applying intelligent capabilities both to network elements and to the entire infrastructure and services. An example is observed in home networks: most complaints arise from incidents in Wi-Fi. Nowadays, AI-based Wi-Fi optimization solutions automatically detect the sources and types of interference and adjust in real-time the transmission power and other parameters, allowing for an increase in Wi-Fi speed of around 20% in environments with interference.
In terms of energy efficiency, previously network equipment operated at full capacity regardless of traffic volume, which generated unnecessarily high energy consumption. Thanks to intelligent traffic analysis, it is now possible to apply dynamic and precise adjustments of power in the OLT chassis and line cards, achieving a reduction in the overall energy consumption of devices by more than 20%.
Regarding intelligent self-repair of faults, the resolution of broadband incidents at home is organized into three levels. Self-repair: problems in the air interface, such as roaming failures, are automatically corrected by O&M AI agents. Remote O&M: incidents associated with configuration parameters are addressed remotely from the NOC with the support of AI agents. On-site visits: problems arising from the physical fiber connection are accurately located with the help of O&M Copilot.
"In the face of the AI era, we must change our philosophy to rebuild the network architecture, reform the business model, and redefine network O&M," concluded Bob Chen. "Through these three 'R's, we can reform the value of the UBB network and move forward together towards new growth in the AI era."
(CONTINUES)