Gain access to resources and project updates

First name

Last name

Business email

Company name

Job title

TM Forum will be processing the above information, with the assistance of our service providers located within and outside the European Union, to manage your registration to this event or report download, as well as to keep you informed about our services and products, future events and special offers, the organization of events, providing training and certification, and facilitating collaboration programs. Privacy policy

I wish to receive further information from the Catalyst Team about their products and service by electronic means. Check the "Team Members" section of this Catalyst Project to review the companies that will receive your information. Companies may join the project in the future, so please check back periodically for any updates

I agree to the tmforum.org Terms and Conditions of Use

All Moonshots

AI Smart CapEx for green and efficient network investments

URN M24.0.665

Topics 5G, AI (Artificial Intelligence), Autonomous networks

Optimizing network rollout autonomously for radio access and fiber

Project companies

Crystalized in an AI-ML product, we bring an in-depth understanding of the telco business and the pain points experienced by CSPs. We focus on easing them in a manner that will maximize acceptance by the different stakeholders. * Building commercial forecasts that map to the different technologies: We start our journey by providing the marketing team with a user-friendly tool that allows them to bypass expensive external consultants and understand the technology adoption trends in the customer base, as well as a systematic approach to handling new use cases for which we lack proper historical data. * Breaking the wall between marketing and network: With marketing focused on customers (and increasingly devices), their forecasts and projections are hard to map into network traffic load, failing to properly align with the Network team's actionable elements, such as the BTS. For this, we provide a systematic approach that makes optimal use of available information to translate high-level commercial projections into street-level traffic load. * Incorporating CX elements into the optimization without overloading the marketing team: Most CSPs have embraced the “Golden sites” methodology that focuses on network elements as revenue-generating. This approach falls short in markets where customer expectations have moved to a more demanding stage, and it fails to acknowledge the impact of differences in network CX across operators on customer retention while tying the network team for several weeks, typically. Incorporating these elements into the solution involves a significant increase in computational complexity that requires an automated approach— essentially extending the concept of Automated Network from Network Operations to Network Planning. Furthermore, our solution targets delivery times that allow testing of multiple commercial scenarios and quarterly, even monthly, reviews. Looking at previous deployments in several countries, AI-ML Smart CAPEX unlocks CAPEX efficiency gains of 15-25% on new Network investments. Additionally, AI-ML Smart CAPEX also enables a 15-25% reduction in OPEX (including electricity consumption) on new CAPEX, along with reasonable spectrum and electricity savings on existing CAPEX— contributing significantly to Net Zero emissions goals by 2030.

Resources

How will the success of the solution be measured? The main benefits of our product are, firstly calculating accurate per-technology traffic forecasts (3, 6, 12, 18, 24,... months down the line) and secondly optimizing the network rollout plan to serve that future demand. To measure success of the traffic forecasts, we track deviations between our traffic forecasts and the actual traffic demanded per site and technology over a period of time. To measure the success of the rollout plan, a naive approach would focus on a comparison of the optimization algorithms, comparing the Smart CAPEX-optimized rollout plans versus rollout plans calculated by other approaches (more simplistics approaches such as Golden Sites or manual rollouts calculated by the CSP’s Network Planning teams). This has the advantage of being an objective, simple to calculate metric. This would be a typical lab measurement, and it would be a typical “engineering” approach. However, we strive to go beyond this and provide a holistic view of how successful our tool and our methodology are. We can not simply focus on the mathematical elements of the solution while ignoring the market environment, we are not operating in a vacuum: Calculating an optimized rollout for a traffic that is not there brings no benefits to the CSPs. We need, firstly, to build an accurate traffic projection, and only then we should focus on how good our optimization algorithms are. Looking at previous deployments in several countries, AI-ML Smart CAPEX unlocks CAPEX and OPEX efficiency gains of 15-20% on the new Network investments and a significant OPEX saving on the existing Network Infrastructure. Additional information https://docs.google.com/presentation/d/1NJD2iQu1PNpQ22aWdJfQiezC6YrWMvlGTxOTfuUC5X0/edit?usp=sharing What is AI-ML Smart CAPEX (Product, Team, Data, Methodology and Case Study)? (slides 1-29) AI-ML Smart CAPEX and ESG. Beyond power savings, how can SmartCAPEX help with ESG targets? (Slides 29-33) Benchmarking different generations of Smart CAPEX (Slides 38-42) Over the time, we have see three main generations of Smart CAPEX Solutions (1) 1st-gen, Network Traffic Analysis, (2) 2nd-gen, Golden Sites and (3) 3rd-gen, Locatium approach In this file we present a benchmark of the different benefits that each approach brings. About Cloud Deployment Smart CAPEX leverages ML and AI algorithms such as Probabilistic search and Monte Carlo simulations on huge datasets (in the range of hundreds of TB) This requires a lot of computation power, in the range of tens of thousands of CPU cores.This is why cloud deployments are ideal. But it pays off as the results are CAPEX savings in the range of 100s of millions of US$ a year for a medium-sized CSP. FBB Generations (slides 43-44) The tool UI (slides 45-46)