2025 Data Center EHSxTech® Insights: Power, Safety by Design, AI, and HRAs

The fall 2025 Data Center EHSxTech® event brought together environmental, health, and safety (EHS) professionals, engineers, and technology leaders from across the data center industry. Hosted on October 8, 2025, at the Salesforce Tower in Dublin, Ireland, the event focused on how artificial intelligence (AI), design innovation, a focus on high-risk activities (HRAs), and smarter energy strategies are shaping the next generation of data center workplace safety.

From AI in EHS, safety by design, training and metrics, and predictive modeling to supplier quality and grid resilience, the message was clear: technology is advancing fast—but human oversight and collaboration remain essential, especially with the rapid growth and expansion of the data center industry.

Below are the top takeaways from this event and actions EHS leaders can apply to data centers today.

1. AI and EHS: From Experiment to Everyday Use 

AI has moved beyond the experimental phase to become a tool used daily to improve data quality, support real time monitoring, and optimize daily workflows. Many organizations are using AI tools like Copilot, Gemini, ChatGPT, and PagerDuty to manage alarms, summarize incidents, and streamline documentation.

One company at the event even created an AI assistant that instantly answers internal safety policy questions, saving hours of research. Others reported using AI for:

• Comparing old and new safety procedures against ISO standards
• Creating incident summaries and action lists
• Drafting training materials and toolbox talks
• Reviewing Risk Assessment Method Statements (RAMS) for gaps
• Using AI scoring to improve the quality of safety observations

But for all of these use cases, governance matters. Organizations are formalizing AI usage policies, privacy guardrails, data-retention rules, and human oversight protocols. Attendees’ AI adoption was by no means uniform, but everyone recognized the importance of incorporating the technology into EHS processes and its potential to raise the bar in the EHS function.

2. Predictive Modeling and Ergonomics: Smarter Prevention 

AI is helping predict risky patterns before incidents happen, especially in high-traffic zones and maintenance areas.

Some real-world examples included:

• Thermal imaging for battery storage areas and electrical safety
• Ergonomics assessments for office and remote workers
• Continuous monitoring of posture and task strain through cameras and sensors

These innovations are helping companies shift from reactive safety to predictive prevention. Participants stressed the importance of piloting small projects, verifying results, and communicating clearly about data privacy and worker consent.

3. Safety by Design: Building Safety into Every Stage 

Designed safety was a recurring theme, emphasizing that forethought and planning for safety need to happen during concept and design phases, not post-construction.

Shared best practices included:

• Having a Safety by Design specialist at 30/60/90% design reviews
• Including technician and vendor feedback in design decisions
• Designing for access, ergonomics, and fall protection
• Using sensors and load monitoring tools for preventative maintenance
• Early incorporation of local regulations, such as worker welfare space requirements

To truly build safety into facilities and processes, design reviews need to go beyond procedure, with purposeful, human-centered attention to detail. Getting input from operations teams is a critical part of eliminating potential hazards before they become problems.

Dive deeper into this topic with our blog “Safety in Design: The First Step in Controlling Workplace Risks”

4. Quality and Supply Chain: Treating Quality as a Safety Issue 

Several discussions brought up equipment and supplier quality challenges like untested cables, transformer failures, and material contamination. Each example led to costly delays, and, in some instances, safety risks.

Solutions that stood out included:

• Conducting Factory Acceptance Tests (FATs) before delivery
• Using owner-side inspectors to visit supplier facilities
• Sharing quality alerts and recalls across teams
• Holding monthly global incident calls to discuss lessons learned

One idea gaining traction was a shared “Quality Alerts Hub” for cross-industry visibility on equipment recalls and supplier risks.

5. Measuring What Matters: Smarter Safety Metrics 

Participants discussed the limitations of traditional metrics, like recordable incident rates, and talked about better ways to measure effectiveness.

Emerging metrics included:

• Percent of high-risk work with verified critical controls in place
• Time to close corrective actions
• AI-filtered observations that prioritize significant findings
• Combining safety and operational metrics (rework, uptime, schedule) for a holistic view

The goal: focus on preventing incidents—not just counting them. Align definitions (like how hours are calculated) and share metrics regularly with leadership and frontline teams so comparisons stay meaningful.

The conversation will continue in coming months, as the Data Center Safety Council publishes calls to action on TRIR limitations within the data center industry.

6. Managing High-Risk Activities 

Organizations are aligning on a common HRA list, consistent critical controls, and testing ways to embed HRA checks in both change management and job safety analysis, not just capital projects.

Nuances added this year:

• Clarifying standard vs. non-standard work scenarios
• Establishing common terms (e.g., “critical controls” vs. “success factors”) to improve clarity
• Real-world examples: critical lifts, live-energy electrical work, vehicle movements, confined spaces

The goal is to establish an effective leading indicator that shows whether the riskiest work is actually controlled, in real conditions.

7. Training That Scales: Digital Inductions and a Global Baseline 

Teams are rolling out digital site inductions, which are short, multilingual video modules with safety culture questions, to save time and standardize onboarding. AI is helping summarize standards into short clips and translate content for global reach.

In parallel, the Data Center Safety Council is advancing a four-hour, self-paced Data Center Safety Awareness Certification to create a consistent baseline for operators and contractors worldwide.

8. Energy and Power: Safety in a Changing Grid 

As data center growth accelerates, energy demand and grid capacity have become major safety and sustainability challenges.

Discussions explored:

• Onsite generation models, including utility-leased/operated plants
• Natural gas as a transitional bridge while capacity catches up
• Battery Energy Storage System (BESS) siting concerns (e.g., electric and magnetic fields (EMF) and community acceptance)
• The benefits of closed-loop cooling to reduce water and environmental risks
• Longer-term options like small modular reactors (SMRs) and carbon capture, with strong community engagement

EHS and sustainability professionals were encouraged to take an active role in:

• Assessing community and environmental risks
• Supporting emergency response planning for new energy systems
• Collaborating with utilities and regulators early in the planning process

9. Operating Across Jurisdictions: Same Outcomes, Local Paths 

Participants compared U.K. and Ireland (stricter enforcement and conservation expectations) with U.S. practices and shared APAC experiences.

The takeaway was that it’s crucial to agree on outcomes and critical controls first, then localize for legal, cultural, and enforcement realities.

Need multi-region support? Check out Inogen Alliance to get support wherever you are in the world.

10. Collaboration Is the Competitive Edge 

Across topics, the winning behaviors were the same: sharing data, aligning definitions, publishing what works, and keeping people in the loop so professional judgment stays sharp.

Key actions for the year ahead: 

• Pilot AI tools for ergonomics, predictive safety, and incident summaries
• Add Safety by Design representation in every major project review
• Launch a global quality-alerts system
• Refine safety awareness certification programs

Looking Ahead 

The future of EHS isn’t just digital; it’s human-led, data-informed, and globally connected.

AI and automation are transforming the world of EHS, but true progress depends on design foresight, supplier collaboration, and a culture that values prevention over paperwork.

Interested in attending our next event? Connect with us to learn more.