Enhancing Network Infrastructure to Support AI Growth: Insights from AWS Event
BOSTON – A recent event held by AWS underscored a pivotal theme: the scaling of artificial intelligence (AI) necessitates more than just advancements in models and data infrastructure—enterprise organizations must also enhance their network capacity. The discussions highlighted the integral role of networks in linking data centers, cloud providers, and field employees, emphasizing a growing need for robust connectivity.
As businesses increasingly embrace AI technologies, the network takes center stage in processing data closer to the point of generation, particularly at the edge. This sentiment resonated with a diverse group of speakers representing various organizations at the Inside AWS Boston event on November 4. Among the attendees were a three-year-old robotics startup heavily reliant on AI, an 80-year-old professional sports franchise selectively adopting AI, and a global telecommunications leader in the midst of digital transformation.
Luke Corbin, the assistant vice president for digital transformation and AI at AT&T, shared candid insights into the challenges posed by outdated infrastructure. "We are a 150-year-old company that has undergone significant mergers and acquisitions," Corbin explained. "This history complicates our data infrastructure, which is cluttered with legacy application components. The challenge then becomes how to construct an integrated infrastructure capable of accelerating AI development. What does your connectivity backbone actually look like?"
According to Gartner’s 2026 Market Guide for Edge Computing Platforms, it is anticipated that by 2028, over two-thirds of enterprise-managed data will be created and processed outside traditional data centers or cloud environments. Furthermore, an IDC study projects that by 2030, half of all enterprise AI inference workloads will occur at endpoints or edge nodes.
Corbin emphasized, "When you handle as much data as we do, and your business relies on it, you must seriously evaluate the data transit between data centers and cloud providers. Do you possess the requisite high-capacity network to effectively manage that?"
The Crucial Impact of AI on Edge Network Upgrades
For AT&T, the reliance on AI agents among field workers highlights the critical need for reliable edge network connectivity to facilitate their tasks. Corbin elaborated, "When you have field technicians or sales personnel depending on this intelligent layer to empower their job functions, they must maintain robust connections, whether through mobile networks or satellite systems, in collaboration with our partners."
On September 8, AT&T revealed plans to integrate Amazon’s Leo satellite network into its architecture for both enterprise and public service connectivity, particularly aimed at providing service in remote areas. Corbin stated, "The importance of this connection cannot be overstated. Missing an email is trivial; failing to connect means you cannot perform your job."
At Luminous Robotics, the necessity for a strong network infrastructure echoes the concerns expressed by AT&T. In this case, the focus is on ensuring that fleets of autonomous robots, responsible for deploying solar energy solutions, can complete their tasks efficiently. The LUMI robots utilize vision-language-action models for real-time edge inference, allowing them to function without tedious reliance on cloud communication.
However, even with these capabilities, Luminous Robotics still transmits all operational data back to AWS for analysis. Jay Wong, CEO of Luminous Robotics, pointed out, "We’ve created a bandwidth-constrained environment. Our operations typically occur in remote locations; deploying solar power plants is not like setting up in downtown Boston."
To ensure effective connectivity, Luminous employs a combination of LTE and Starlink network infrastructures. Tasks that require high transmission frequency, such as error logs or safety alerts, depend on a high-speed LTE network, while Starlink services less frequent, high-bandwidth transmissions.
As the company continues to evolve, improving network capacity will be a significant focus. Wong shared, "We have plans to expand our manufacturing capabilities to produce hundreds of robots. Currently, we operate in an 11,000-square-foot facility, and as we scale our operations, networking constraints will come to the forefront."
Wong remains open to exploring future partnerships, including potential collaborations with Amazon to bolster networking infrastructure: "Our needs for our robots can be met by various partners, so we are agnostic in that regard."
Network Security Measures in Sports
The Boston Celtics also find themselves needing to expand their network infrastructure as their organization experiences rapid growth. Chief Technology Officer Jay Wessland noted that increased data from both on-court activities and off-court endeavors has necessitated a scaling-up of their technological setup. "Every game is broadcasted through multiple channels, and stats can be relayed through several platforms," he stated.
In this evolving digital landscape, the Celtics utilize various cybersecurity tools at both the network and user levels. Wessland emphasized the integration of ReliaQuest’s GreyMatter, a security platform that employs AI to sift through security data and escalate alerts for human review when necessary. He elaborated, "Our system not only auto-generates information concerning network packets but also flags potential issues for scrutiny by our third-party partners, leading to significant workload management at the network level."
In conclusion, as organizations pivot toward more advanced AI applications, the underlying network infrastructure will play an essential role in facilitating this technology shift. From telecommunications to robotics and sports, improving connectivity is increasingly becoming a primary concern for enterprises invested in harnessing the power of AI.