TL;DR

  • AI infrastructure development is accelerating, but power availability, site selection, and permitting challenges are influencing where and how new data centers are built.
  • Growing demand for AI training and inference is driving innovation across compute, networking, memory, and storage infrastructure.
  • Delivering AI capacity at scale requires greater coordination between infrastructure design, energy planning, capital investment, and operational strategies.
  • As development expands, political considerations, regulatory requirements, and local community expectations are becoming increasingly important to project viability.

# # #

The rapid expansion of artificial intelligence (AI) is transforming digital infrastructure, creating new demands for computing capacity, power availability, networking, and data center development. These challenges were central to discussions at AI Infra Summit 2026, held September 15–17 in Santa Clara, California, where industry leaders explored the technologies, investment strategies, and operational considerations shaping the next generation of AI infrastructure.

Bringing together executives, technology innovators, infrastructure developers, and investors, the summit examined the increasingly interconnected systems required to support AI at scale. From high-performance computing and data movement to power constraints, site selection, and infrastructure economics, the discussions reflected an industry working to balance accelerated demand with the practical challenges of building and operating AI-ready environments.

Power Availability and Site Selection Are Reshaping Development Strategies

As AI workloads drive demand for larger and more energy-intensive data centers, securing viable development sites has become increasingly complex. Power availability, utility interconnection timelines, permitting requirements, and local political considerations are influencing project feasibility long before construction begins.

These challenges were central to discussions moderated by Ilissa Miller, CEO of iMiller Public Relations, who joined industry leaders examining the intersection of power, permitting, politics, and site selection. The discussion featured Raul Martynek, CEO of DataBank, and explored the factors influencing the race to secure suitable land for AI infrastructure development.

The growing importance of early-stage planning reflects a shift in how developers evaluate potential projects. Access to land and capital alone is no longer sufficient. Successful development strategies must also account for power delivery, regulatory approvals, infrastructure readiness, and the concerns of communities where facilities are proposed. These considerations are becoming essential to reducing project delays and improving long-term viability.

Scaling AI Compute Requires a More Integrated Infrastructure Approach

The increasing complexity of AI applications is placing new demands on computing architectures and the systems supporting them. As organizations expand their use of large language models, agentic AI, and real-time inference, infrastructure strategies must address performance, scalability, and efficiency across the entire technology stack.

The evolution of AI computing was explored by Ian Buck, Vice President, Hyperscale & HPC Computing at NVIDIA, whose participation in the summit included discussions on advancing infrastructure for agentic AI and the technologies supporting frontier AI development. These topics reflected the growing need to align processing capabilities with the networking, memory, and storage resources required to support increasingly sophisticated workloads.

The summit also highlighted the importance of designing infrastructure around specific AI requirements rather than relying on conventional computing models. As training and inference workloads evolve, organizations must consider how specialized hardware, distributed computing, and more efficient system architectures can improve performance while managing the cost and complexity of deployment.

Data Movement and Networking Are Critical to AI Performance

While computing capacity remains central to AI infrastructure development, the ability to move and access data efficiently is becoming equally important. Large-scale AI environments depend on high-bandwidth, low-latency connectivity between processors, memory, and storage systems, creating challenges that cannot be resolved through additional compute resources alone.

Among the experts addressing these challenges was Nick Harris, Founder and CEO of Lightmatter, who participated in discussions examining data movement, memory, and storage bottlenecks at scale. The session reflected the growing importance of improving how information moves throughout AI systems as organizations pursue greater performance and infrastructure efficiency.

Networking architecture, optical connectivity, and data transfer efficiency are increasingly important considerations for AI infrastructure investment. As distributed computing environments expand, improvements across these systems will be essential to supporting larger workloads, reducing performance limitations, and maximizing the value of computing resources.

Building AI Infrastructure for Long-Term Operational Efficiency

The rapid pace of AI infrastructure investment is creating pressure to deliver new capacity quickly while maintaining reliability, flexibility, and cost control. As facilities become larger and more technically complex, operators must evaluate how design decisions affect long-term performance, energy consumption, and the ability to adapt to future workloads.

The importance of integrated infrastructure planning was addressed by Ram Nagappan, Vice President, AI Infrastructure and Lead Architect, Gigawatt-Scale Data Centers at Oracle, whose presentation focused on the design of AI campuses at gigawatt scale. The topic highlighted the need to coordinate power infrastructure, computing requirements, and facility architecture when planning increasingly large AI deployments.

These considerations extend beyond initial construction. Operational strategies must account for changing power demands, evolving hardware requirements, and the integration of systems that support continuous, high-performance computing. Greater coordination between infrastructure developers, technology providers, and operators will be essential to managing complexity and supporting sustainable growth.

Building the Infrastructure Behind AI’s Next Phase

AI Infra Summit 2026 highlighted the scale of the challenges and opportunities facing the digital infrastructure industry. While advances in computing technology continue to accelerate AI adoption, the ability to support that growth depends on coordinated investment across power, networking, data center development, and operational systems.

The discussions reinforced that AI infrastructure cannot be planned or evaluated in isolation. From securing viable development sites to addressing data movement bottlenecks and designing gigawatt-scale campuses, the industry’s next phase will require closer alignment between technology innovation, infrastructure investment, and the communities supporting new development.

As demand for AI continues to expand, the strategies explored in Santa Clara provide a framework for addressing the practical requirements of scalable, resilient, and efficient digital infrastructure. Industry leaders will have another opportunity to continue these discussions at AI Infra Summit 10X 2027, scheduled for August 30 – September 2, 2027, in San Jose, California.