Breakfast provided at the conference area of the hotel.

Buses will depart from the front of the hotel lobby for attendees that signed up for the tour. Security check-in will be conducted upon arrival at the site. Photo IDs are required for security check-in.

Attendees will participate in a guided tour to explore the site’s infrastructure, security protocols, and operational practices.

The bus will transport attendees back to the hotel following the conclusion of the tour.

Tour attendees will meet for a working lunch to note best practices and opportunities for improvement observed during the data center tour.

In this AI Working Group session, senior IT leaders will engage in a candid, peer-driven discussion on the realities of AI implementation—balancing infrastructure demands, operational risks, and evolving regulatory landscapes. Building on insights from past sessions in Portland and Paris, this roundtable will explore practical strategies, roadblocks, and innovations shaping the AI-driven enterprise.

The Green Grid (TGG) has published a methodology and supporting server work capacity values by CPU Part Number that will enable operators to calculate Cserv for standard CPU servers with Intel, AMD, IBM Power and selected ARM processors. Uptime has proposed a simplified method to calculate Cserv for GPU based servers and Cstor for storage products. This workshop will discuss the methodology for calculating ICT capacity for server and storage products, provide example calculations, and answers questions about the calculation methodologies.

Official welcome and introductory remarks to commence the conference.

• Andy Lawrence, Executive Director of Research, Uptime Institute
• Chris Brown, Chief Technical Officer, Uptime Institute
• Jay Dietrich, Research Director, Sustainability, Uptime Institute

Uptime Institute presents the popular Panorama session, where the panelists and delegates discuss trends, events, and outages affecting the digital infrastructure industry.

An opportunity for attendees to network over dinner and drinks.

Breakfast provided at the conference venue.

Panel Moderator: Chris Brown, Chief Technical Officer, Uptime Institute

By 2030, facilities will need to be more adaptable, autonomous, and resilient than ever before. This discussion dives into the key design and construction shifts necessary to meet these challenges—from modular scalability, to AI-workload densification, to climate risk mitigation and supply chain resilience.

Panel Moderator: Matt Stansberry, Executive Director Data Center Practitioner Insight & Experience, Uptime Institute

Historically, equipment lifecycle planning involved tracking the normal wear and tear against assumptions on the Expected Useful Life (EUL) of the equipment. Now, organizations face rapid technological changes and early obsolescence at a component level, disrupting conventional planning processes.

Time allocated for attendees to take a break and network.

Presenter: Vladimir Prodanovic, Principal Program Manager, NVIDIA

This session will provide an update from NVIDIA on the latest trends for implementing AI clusters, as well as a look ahead at how AI will be implemented in the future.

Presenter: Max Smolaks, Research Analyst, Uptime Intelligence

AI is not a single, uniform workload. The infrastructure needs of a model depend on its application, the stage of the model's lifecycle, performance levels required, and other factors. Delivering AI at scale will require a variety of hardware platforms and architectures. In this session, we will look at the state of the AI accelerator market in 2025, and observe the different directions taken by hardware designers.

Presenter: Owen Rogers, Senior Research Director, Uptime Intelligence

As AI adoption accelerates, organizations face critical decisions about where to train and deploy their models.

• Should you leverage the flexibility of the cloud or invest in dedicated infrastructure for long-term cost efficiency and business value?
• How are AI models trained and then used in production?
• How does utilization affect the TCO of cloud vs dedicated infrastructure?
• How do diminishing returns impact the merits of different approaches?

Lunch break for all attendees.

Presenter: Jay Dietrich, Research Director, Sustainability, Uptime Institute

This panel discussion will focus on the opportunities for enterprise operators to improve the efficiency of their IT infrastructure, and for colocation operators to work with their IT tenants and customers to drive IT efficiency. The result is improved overall IT efficiency for the tenants and business opportunities for the colocation operators.

Presenter: Mohamad Hnayno, Doctor of Engineering in Thermal and Energy, R&D Engineer for OVHcloud

This presentation highlights an experimental study conducted by the OVHcloud lab on an innovative single-phase immersion liquid cooling technique. The approach combines direct-to-chip water cooling with passive single-immersion cooling. The study evaluated the impact of three different dielectric fluids, multiple water circuit configurations, and varying server power profiles. The presentation concludes with a comparative analysis of cooling performance across air-cooled, liquid-cooled, and the new hybrid system.

Moderator: Jacqueline Davis, Research Analyst, Uptime Intelligence

Liquid cooling commissioning is one of today's most daunting data center facilities challenges as organizations begin deploying these next-generation systems. Many organizations struggle to employ the rigorous testing practices required to ensure systems function as designed.

Short recess for attendees.

Presenter: Daniel Bizo, Research Director, Uptime Intelligence

As IT rack power demands rise and space constraints tighten, data centers will likely be forced to rethink their approach to power distribution. One option on the horizon is bringing medium voltage distribution closer to the white space. However, the attractiveness of that will likely depend on related technical innovations, such as solid- state transformers. This presentation explores the driving forces behind as well as the hurdles in the way of introducing higher voltage levels in data centers power protection and distribution.

Moderator: Nick Archer, Regional Technical Director, Europe & UK, Uptime Institute

An outage is never just about the moment it happens—what comes next is just as critical. A well-executed post-outage response can mean the difference between a full recovery and lingering risks. This panel will dive into the essential policies, best practices, and real-world lessons learned from handling outages. Topics include structured post-mortems, incident documentation, stakeholder communication, compliance considerations, and how to turn failures into future resilience.

Andy Lawrence, Executive Director of Research, Uptime Institute Ali Moinuddin, CCD & CSO, Uptime Institute
Antonio Ramos, President & CEO, LEET Security
John O’Brien, Sr. Research Analyst, Uptime Institute

In an era where cyber threats, physical security risks, and operational vulnerabilities are becoming more disruptive, data center cybersecurity still lacks urgency. This session will test the seven most common fallacies still held, including why zero-trust policies and physical air gaps are not always practical, and why ransomware attacks are a real risk. Plus, includes discussion of recent security/resiliency legislation.

Moderator: Scott Killian, VP Global Uptime Network Membership Services

An open forum for attendees to engage with experts and peers on pressing issues in the data center industry. Participants are encouraged to share insights, pose questions, and discuss innovative solutions in a collaborative learning environment.