AI to the next level
Key Highlights
- AI is now the top priority at the Fiber Broadband Association, driving demand for significant fiber infrastructure growth, including 213 million additional fiber miles for data centers.
- Streamlined permitting processes and increased funding are essential to accelerate fiber deployment and support the AI revolution across communities and industries.
- Physical AI applications in agriculture, manufacturing, and construction are improving efficiency, safety, and sustainability through real-time data collection and edge computing.
- Investments in fiber, IXes, and long-haul routes are critical for lowering costs, enhancing network resilience, and supporting AI data center expansion nationwide.
- Policy initiatives like the BEAD Act aim to leverage federal funds for infrastructure and workforce development, ensuring robust connectivity to unlock AI's full potential.
By Gary Bolton / Fiber Broadband Association
At the Fiber Broadband Association (FBA) Premier Member Meeting in December, Artificial Intelligence (AI) emerged as the most critical topic of discussion for attendees and companies. Out of a straw poll of 12 topics affecting FBA membership, “AI Infrastructure & Emerging Technologies: New Business Models for Operators” ranked number one, surging ahead of issues such as BEAD implementation and supply chain and permitting reform by a substantial margin.
Everyone sees opportunities in AI and is highly focused on what AI can mean for their business. From a fiber perspective, AI data centers will require an additional 213 million fiber miles, according to the most recent research published by RVA and FBA. The number includes 109 million miles for route upgrades, 66 million to support new data center construction, and 38 million more miles for new long-haul routes. For security and resilience, this fiber should be deployed underground and across redundant paths. It should leverage the medium’s unique ability to serve as a sensor to detect attempts to cut or wiretap it.
To achieve this growth and meet our country's infrastructure needs, we will need a holistic approach that addresses the pain points that slow construction, along with a combination of public and private investment to ensure sufficient fiber is built to support the AI revolution and the data center construction boom.
Permitting has been, and continues to be, a challenge for rapidly deploying fiber. The FBA and others have developed best practices and models that municipalities can adopt to streamline and facilitate the permitting process. Still, funding is needed for local governments to invest in cloud-based permitting management solutions and the personnel to operate them. Having cloud systems easily accessible to both vendors and the public improves transparency and maximizes efficiency compared with paper or legacy back-office software solutions. I am strongly encouraged by the efforts on Capitol Hill to apply BEAD non-deployment funds to improve permitting at the local and state levels.
The effort to enable the use of BEAD non-deployment funds for other network infrastructure improvements is welcome and strongly encouraged. Adding more middle-mile long-haul routes and fiber provides numerous benefits to local communities and the country at large, including lowering the overall cost of network operations through competition and building a more resilient network that is less vulnerable to single-point-of-failure disruption.
Building more internet exchanges (IXes) across the country outside of highly dense rural areas would also be a welcome and wise investment, helping to lower broadband connection costs and network latency for many mid-sized and rural municipalities, a move especially needed as AI data centers are increasingly built in areas where land is cheap. There is available power, or it can be built quickly to meet the needs of new construction.
From AI to AI
Building AI will require more than fiber and software; it will also require a change in mindset. We need to think less about the generic marketing hype of Artificial Intelligence and focus more on Applied Intelligence: how to apply the infrastructure, tools, and systems being created and scaled today to provide solutions that increase productivity, improve safety, health, and well-being, and provide a clear return on investment.
Applied Intelligence is already in everyday use for recognition tasks, such as voice, face, and shape, enabling hands-free voice interfaces in our homes and motion alerts on our doorbell and security cameras. But AI has steadily expanded to include physical systems, leading to a physical AI category encompassing robotics, manufacturing, and autonomous vehicles.
Industrial AI integrates engineering knowledge and experience production to create what Siemens calls “adaptive production.” Leveraging automation, AI, and digital-twin technology, adaptive production enables real-time responsiveness to changes or disruptions. Manufacturers can rapidly adjust to demand fluctuations, customize products for local markets, and respond to supply chain and other issues.
Agriculture and construction take charge
Across the agriculture and construction industries, companies such as Caterpillar, Daewoo Bobcat, John Deere, and Oshkosh are melding heavy vehicles with AI to improve operations, reduce manpower and energy used, and increase productivity. Physical AI vehicles in the fields and quarries collect and transmit data from the real world in real time to historical cloud-based AI models, providing detailed insights into crop and ore quality and how to harvest them correctly.
Precision agriculture is already making a difference at a New Zealand cherry farm. The New Zealand Cherry Corporation is sending over 400,000 HD photos per hour over its 25G PON connection during growing and harvest season over a 25G PON connection, enabling the company to rapidly evaluate the status of the crop in real time and optimize when and where it should harvest. Better cherries mean more profits instead of rejected fruit. The business anticipates nearly tripling its harvest with the new technology.
Closer to home, Land O’Lakes is leveraging AI through a strategic partnership with Microsoft to help its 10,000 co-op members be more sustainable, more productive, and more profitable. The alliance will focus on co-developing AI-powered tools including a digital assistant called “Oz.” Farming has always been a data-driven business, with information collected at the field and farm level, but 21st century AI enables the creation and use of models to provide recommendations to farmers to make smarter decisions on when to apply water and other inputs based on soil types, environment, and current weather, with information collected from IoT devices, vehicles, drones, and satellite imagery.
This complicated symphony of precision agriculture and physical AI is illustrative of the broader trend where edge AI devices and processes perform complex and latency-sensitive tasks and then send the results upstream into one or more cloud-based AI processes to derive insights and provide recommendations for actions, which then flow back down for evaluation and implementation by the end-user. In addition, collected information is archived, added to larger models, and used to refine and update training for edge processes in the cloud, with revised versions shipped when ready.
We can and will continue to see this model play out across industries such as transportation, mining, construction, manufacturing, and energy production, with the goal in most cases to improve the productivity and safety of the existing workforce by having machines perform repetitive and dangerous jobs that are difficult to fill today, rather than replacing experienced workers and their specialized skillsets.
Continued investment in fiber and supporting infrastructure, such as IXes, is critical to unlocking AI’s full potential nationwide. Building and connecting AI data centers is only one step in the journey, as increased connectivity will be needed to enable mobile, home, and business users to access those resources and to provide the information they need to perform their tasks.
On Capitol Hill, the Supporting U.S. Critical Connectivity and Economic Strategy and Security (SUCCESS) for BEAD Act outlines the use of remaining BEAD funding to strengthen and sustain broadband deployment projects, with the option for states to invest in both advanced telecommunications infrastructure and the workforce necessary to deploy it to scale AI. I am confident that the public and private sectors will step up to invest in the critical, ongoing fiber infrastructure to drive AI’s growth and sustain economic success.