As NJFX prepares the first floor of our cable landing station and data center campus for a 10 megawatt AI-ready data hall, it is essential that we engage directly with the leaders shaping the future of this technology and infrastructure.

Artificial intelligence is no longer experimental. It is already influencing how healthcare is delivered, how traffic systems are managed, how security environments are monitored, and how critical systems operate in real time. The pace of adoption is accelerating, and infrastructure must evolve just as quickly. For operators, this means moving beyond assumptions and gaining firsthand understanding of what these workloads require.

That is why our recent visit to NVIDIA’s headquarters in San Jose was so valuable.

NVIDIA continues to set the pace in accelerated computing, delivering advancements on a consistent annual cycle that is reshaping expectations across the industry. The next generation Rubin GPU platform represents a significant shift in how AI systems will be deployed. As power densities increase, these platforms are designed to operate in liquid-cooled environments, placing new demands on facilities that intend to support them.

Being on site allowed our team to spend time with NVIDIA engineers discussing electrical design, cooling approaches, and operational considerations tied to these next-gen systems. We were able to inspect racks populated with GB200 and GB300 platform clusters and observe active live lab environments where new clusters are being tested prior to market release.

Observing these systems operating in real conditions provides clarity that only hands-ons knowledge can provide.  AI infrastructure is no longer an extension of traditional data center design. It introduces structural, mechanical, and operational challenges that must be addressed early in the planning process.

Challenges that traditional carrier hotels and legacy data center designs will face will be significant. For instance, raised floor environments are not well suited for modern AI racks due to structural load constraints and the weight of high-density systems. As rack weights continue to increase, slab-based designs with reinforced foundations become essential.

Liquid cooling further amplifies the importance of detailed planning. Drainage strategy and leak mitigation becomes a foundational design consideration, including how water is routed, how much volume the facility must accommodate, and how containment is implemented. Questions such as curb height, water containment zones, and leak mitigation. They are core elements of building a safe, resilient, and operational AI-ready facility.

“Seeing the equipment live was extremely valuable from an operations standpoint. Reviewing the liquid-cooled manifolds and how cooling is delivered from multiple sources gave us a much clearer understanding of the operational demands these systems place on a facility,” said Brad Graves, Infrastructure Manager at NJFX

Attention to these details can determine whether a facility is able to support advanced workloads reliably over time. Visits like this help surface the smaller but critical design decisions that often become make-or-break factors in high-density AI deployments. They allow us to ask the right questions before construction begins and to evaluate best practices based on real-world performance rather than assumptions.

“The NVIDIA visit highlighted how important flexibility has become at the rack level. Busways allow us to adapt power distribution around high-density AI racks, which is essential as these platforms continue to evolve,” said Ryan Imkemeier VP of Operations at NJFX

For NJFX, this level of diligence is essential as we prepare our 10MW data hall. Our position as a carrier-neutral cable landing station places us at the intersection of global connectivity and emerging compute demand. As AI workloads move closer to data sources and international networks, the alignment of power, cooling, structure, and connectivity becomes increasingly important.

Engaging directly with technology leaders, vendors, and engineering teams ensures we are making informed decisions for both NJFX and our customers. We are taking a deliberate approach, educating ourselves on proven designs, selecting the right equipment, and building infrastructure that is aligned with where AI technology is going.

NJFX will continue to strengthen our partnerships with the company’s driving innovation while actively educating the broader market on what it truly takes to support next-generation AI workloads. Sharing lessons learned, aligning on best practices, and building infrastructure with intention are critical to ensuring long-term success for the