Can Edge & 5G Keep Up With AI’s Demands? What Telcos & Data Companies Think

Artificial intelligence (AI) is a demanding beast, from a data processing point of view.

And yet, while we are used to pinging data server farms to pose our demands to ChatGPT and other AI writing tools, new multimodal tools are emerging, like Google Gemma, that aim to bring AI on-premises, for bespoke purposes, perhaps tailored to individual companies.

With AI needing low latency to perform, the next frontier looks to be moving away from the cloud and towards the edge, as well as needing reliability from 5G networks.

Techopedia talked to executives of leading telcos and data companies to understand the changes needed, identify challenges and opportunities, and to provide guidance and advice for IT managers and business owners as the new AI-mobile migration and transformation begins.

This shift, while it may appear to be natural and seamless, is challenging. The underlying edge infrastructure and 5G’s role in enabling innovation and technology is often overlooked. To support the ever-demanding processing power that AI imposes, the edge and 5G industry must reengineer, adapt, and enhance global infrastructures.

Table of Contents Table of Contents

1. Key Takeaways
2. The Opportunity of AI and Automation
3. Edge Engineering
4. The Need for a New Edge
5. Business and Innovation Opportunities
6. Building The New AI Edge 5G
7. The Bottom Line
8. References

Show Full Guide

Table of Contents

1. Key Takeaways
2. The Opportunity of AI and Automation
3. Edge Engineering
Show Full Guide
4. The Need for a New Edge
5. Business and Innovation Opportunities
6. Building The New AI Edge 5G
7. The Bottom Line
8. References

The Opportunity of AI and Automation

Many experts believe that AI — the same technology that is putting pressure on the current edge and 5G infrastructures — can be used to make this new edge 5G transformation more efficient, rapid, and seamless while reducing costs and resources.

Puneet Sethi, SVP of Product Management and Engineering Operations at private 5G provier Ataya talked to Techopedia about the role of AI and automation in this transformation.

“We believe private 5G and mobile networks will continue to be more automated. AI models will help develop better insights for troubleshooting and monitoring of network and user experience, and help with root cause analysis in case of an issue.”

Sethi explained that generative AI models can simplify the provisioning and configuration of complex networks. But he warned that as more of the AI processing is pushed to the edge, hardware changes will be required.

Edge Engineering

While AI and automation can help develop, design, code, deploy, operate, and monitor new 5G and edge networks, the edge computing structure itself must undergo changes. Ahmed from NTT spoke about the issue.

“Edge computing devices and infrastructure will need to be deployed closer to the end-users to reduce latency and support AI applications. This means edge resources will play a critical role in making sure business goals are met.”

The NTT report Edge Advantage found that eight out of 10 enterprises expect their dependency on third-party edge services to grow over the next two years.

Shahid Ahmed, Executive Vice President of the Group New Ventures and Innovation of NTT — a global technology company and leader in mobile network innovation — talked to Techopedia about what this AI-edge transformation implies.

“The transition of artificial intelligence from cloud-based systems to mobile networks and 5G will necessitate a significant infrastructure transformation that we are already seeing today as more companies adopt a forward-looking edge strategy.”

Ahmed explained that this transformation requires an accelerated adoption of 5G networks to provide the high-speed, low-latency connectivity required for AI applications and IoT deployments.

The new infrastructure will enable a wide range of AI-powered smart devices powered by compute resources closer to where data is being collected, Ahmed explained.

“Due to latency, the cloud will not be able to keep up with the demands generative AI will place on the infrastructure we have today. Businesses need the right expertise to ensure their legacy systems and protocols don’t get in the way of complex solutions.”

NTT recognizes that while AI processing will increasingly move to the edge, cloud infrastructure will still play a crucial role in training AI models, storing large datasets, and orchestrating AI workloads across distributed environments.

“Seamless integration between edge and cloud infrastructure will be essential to support hybrid AI deployments,” Ahmed added.

The Need for a New Edge

IT consulting company Accenture‘s report “Leading with edge computing: How to reinvent with data and AI” revealed that 83% of leading executives from different countries believe that edge computing will be essential to remaining competitive in the future. 81% of executives also said that failure to act quickly “can lock them out from the full benefits of the technology.”

Accenture’s report concluded that edge computing is poised to accelerate innovation leading to new revenue opportunities for companies that evolve by integrating the power of the cloud, data, and AI with essential edge innovation to accelerate the delivery of differentiated experiences. The study shows that approximately only 30% of companies surveyed have a deeply integrated edge with their digital core.

But working in edge infrastructure involves a set of unique problems.

While data centers are built indoors or in controlled conditions and environments, edge centers are not. Besides the technical capabilities that edge hardware and edge centers need to have to support AI workloads, they must also endure harsh conditions.

Edge hardware deals with high and low temperatures, must deliver despite environmental conditions, must be shock and vibration-resistant, and work with limited resources in isolated or rural areas.

Nathan Blom, Chief Commercial Officer of data centre cooling company Iceotope explained that far-edge computing requires high-performance computing, which generates more heat than traditional servers. Air-cooled systems cannot handle the increased heat output, which can lead to issues with server performance. Blom spoke to Techopedia about extreme-edge deployments.

“Extreme edge deployments tend to have reduced control over environmental exposure during installation, operation, and maintenance compared to centralized data center facilities.”

Blom added that when designing their cooling system, they considered the demands imposed by harsh edge deployments and the importance of providing reliable data processing in the face of challenges that range from power constraints to service accessibility, as well as local environment and ambient weather conditions.

Similarly, besides cooling systems — which are vital for edge performance, sustainability, and higher processing — all other edge infrastructure components will also have to meet these demands. Bom explained how the data center’s role is evolving and the edge is expanding.

“High-performance computing, artificial intelligence, IoT, and 5G are transforming the way we do business. As a result, the data center’s role is evolving with more processing being done at the edge where data is created. Now more than ever, edge devices must be reliable and easy to service to maintain uptime and ensure optimal performance.”

Edge: Energy and Sustainability

Sustainability on the edge is not just necessary to reduce carbon emissions in extreme edge locations or rural areas, it is a hardware imposition.

Blom from Iceotope explained that telco deployments that will support AI on the edge will encounter, in some locations, a limited amount of available electricity for individual extreme edge sites.

“The wires that are literally pulled over the poles can reach their destination, but they just don’t have enough electrons inside them to do what it is that we’re asking of our 5G networks (and eventually, our 6G networks).”

Blom explains that the focus should be on reducing energy consumption and greenhouse gas emissions as well as carbon emissions associated with traditional power and cooling processes at the edge.

“If we can address that in a ruggedized and modular way, it means we can be scalable outside on the extreme edge.”

Business and Innovation Opportunities

The Grand View Research report valued the global 5G edge computing market at an estimated size of $1.90 billion in 2022, growing at a CAGR of 49.8% from 2023 to 2030.

The report highlights 5G edge computing as fundamental to support innovation and bring computation and storage capabilities close to the end users and devices including smartphones, IoT devices, and autonomous vehicles. The report adds that 5G edge computation is also required for high-bandwidth and high-reliability applications, such as augmented reality, virtual reality, gaming, and industrial automation, all sectors experiencing growth.

However, as Ahmed from NTT explained, while AI processing will increasingly move to the edge, cloud infrastructure will still play a crucial role in training AI models, storing large datasets, and orchestrating AI workloads across distributed environments.

“Seamless integration between edge and cloud infrastructure will be essential to support hybrid AI deployments.”

Cloud-AI-edge-5G — all-in-one — solutions are in demand. On February 26, NTT and Schneider Electric, announced the beginnings of Private 5G (P5G) — or effectively on-premise private networks which bring together the cloud, edge, private 5G, IoT, and modular data centers to provide edge connectivity anywhere, even in the most remote locations.

NTT recently deployed the technology in Marienpark, Berlin, Germany, and has deployed private 5G RAI Amsterdam, the City of Las Vegas, North Carolina, and for companies and smart factories such as Schneider Electric and the BMW Innovation Hub.

Ataya is also investing in this business model with solutions targeting private 5G deployments of large and small sizes. Their new products Harmony and Chorus come with a cloud-enabled dashboard that relies on AI-driven insights to better manage the network.

Building The New AI Edge 5G

Ahmed spoke about deploying and managing edge computing infrastructure requires significant investment and coordination.

“Consistent and secure edge application performance depends on your organization’s network connectivity. This means that future edge deployments strongly correlate to a campus network overhaul. Also, enterprises cannot neglect network connectivity that is capable of supporting the needs of generative AI on the edge.”

Ahmed said that leaders need to consider that an edge deployment also requires tight orchestration of hardware, platforms, systems, and devices; consistent operational performance without compromising security; and overcoming legacy infrastructure and technical debt.

These and other requirements are the reason why partnership and collaboration among the supply chain and different organizations and companies is essential.

For example, Iceotope´s latest technology is born from a partnership with Intel and Hewlett Packard Enterprise (HPE). They developed KUL RAN, a liquid-cooled server ideal for the extreme conditions of telecom edge deployments.

In contrast, NTT partnered with Qualcomm Technologies to deliver 5G-ready devices with Qualcomm Technologies’ 5G chipsets, which integrate AI models to enhance AI at the edge.

Ahmed explained the role of third-party providers and the importance of being open to collaboration.

“A third-party edge service provider can step in, support the organization with the expertise they need most, and deliver a secure, reliable, low-latency network to support the movement of data and future edge deployments.”

The Bottom Line

From new edge and 5G hardware and software to deal with extreme locations, updating the global edge of telco networks demands collaboration between different sectors, investments, and innovation.

Generative AI took the world by storm overnight, but it still mostly operates in the cloud, limiting its potential. As AI moves and expands through edge and 5G networks, industry leaders focus on scalable, secure, compliant, and sustainable solutions.

With 6G on the horizon and AI getting better by the minute, the AI edge 5G transformation can not be a once-and-done event but will be ongoing, and will continue to adapt and grow to overcome emerging challenges that new technologies create.