Data centers face a simple problem that keeps getting harder: they need more power, but getting that power takes time. The 8th Time-to-Power & Capacity for Data Centers Summit brought that challenge into focus in Dallas on 18–19 August 2026. AI growth, grid queues, permitting delays and rising capacity needs all put pressure on new projects. The programme therefore looked beyond the grid connection itself. It explored onsite generation, batteries, cybersecurity, AI load planning and other practical ways to bring power to data centers faster.

Exploring the road to faster power

The discussions covered different parts of the same problem. How can developers find suitable power markets? How can they avoid delivery delays? And what happens when AI workloads change faster than infrastructure plans?

Day one

Grid Under Pressure: Managing Data Center Growth and Capacity Constraints in High-Demand Power Markets

José RODRÍGUEZ – Chief Operating Officer, New Era Energy & Digital

Finding power is no longer just about finding a suitable site. This session looked at ERCOT, PJM and MISO through the lens of time-to-power and long-term scalability. It also examined who pays for the substation when infrastructure financing falls short. Meanwhile, the discussion considered the trade-offs between congested Tier 1 markets and secondary or tertiary regions.

How We Secured Power for a Large Data Center Project: Timeline, Bottlenecks and Lessons

Speaker not specified in the programme

This session focused on what actually happens when a large data center project needs power. The discussion compared planned and actual energisation milestones. It also looked at transmission limits, transformer lead times, utility queues and permitting delays. More importantly, it highlighted coordination between utilities, EPCs, grid operators and data centers. Developers, hyperscalers and colocation providers could apply those lessons.

The Devices Between You and Downtime

Joel GOINS – Director – Technical Account Manager, DRAGOS

A data center relies on far more connected equipment than its IT teams may realise. This session looked at UPS systems, PDUs, cooling equipment, BACnet controllers, rack monitors, cameras and access controls. It also examined five areas of device security: asset inventory, credentials, firmware, configuration and end-of-life risk. For example, one major data center network achieved complete xIoT discovery in about 12 minutes.

From Grid Consumer to Regulated Entity: The Compliance Horizon for Data Centers

Markus MUELLER – Field CISO, Nozomi Networks

Data centers are becoming major participants in power systems. That shift could bring greater regulatory oversight. The session used FERC, NERC and NIS2 as examples of where data center compliance could head. It also separated today’s requirements from future expectations. At the same time, the discussion stressed stronger visibility, cybersecurity and operational controls. Companies that prepare early can avoid expensive retrofits driven by future audits.

Solving Time to Power: Soluna’s Approach to Accelerating Data Center Interconnection

Nick LANCASTER – VP of Engineering, Soluna Holdings

AI computing continues to put pressure on available power. This session examined how data center design and Soluna’s behind-the-meter model can shorten the route to power. It also drew on lessons from Projects Dorothy and Kati. In practice, the discussion connected the wider power bottleneck with specific approaches to interconnection. It gave developers another way to think about how project design can affect delivery timelines.

Bringing Connectivity Risks in Data Center Facilities to the Surface

Sean IR – Director of Product Marketing, Claroty

Connectivity creates another risk that data center operators cannot ignore. This session examined how attackers can target the operational backbone of a facility. It also looked at the visibility gaps that leave assets exposed. From there, the discussion moved toward network resilience and auditable compliance. The focus stayed practical. Operators need to know what connects to their infrastructure before they can properly understand the risks around those connections.

Load Flexibility Using BESS

Uzair BAIG – Head of Power Strategy, Iron Mountain

Battery storage can do more than provide backup power. This session explored how behind-the-meter BESS can turn large data center loads into more flexible grid assets. The discussion used Iron Mountain’s Edison, New Jersey facility as a case study. The site features a 12MW BESS and a large rooftop solar array. It also demonstrated tariff neutrality and opportunities to capture transmission, capacity, and ancillary-market value through the PJM value stack.

The deteriorating PR situation for building new data centres in the US

Giovanni CAPRIGLIONE – Texas State Representative HD98, Texas House of Representatives

Power availability does not settle every challenge around a new data center. Communities also want to know what a project brings to the area. This session suggested leading with quantified community benefits instead of corporate defence. It also called for formal industry coalitions and early publication of hard data. In addition, it recommended using a trusted Texas spokesperson who can speak credibly to local communities.

Planning for the Unknown: Power Forecasting in the Age of AI Volatility

Nick LANCASTER – VP of Engineering, Soluna Holdings
Ramiro MARTINEZ, PE – AVP Data Center Operations, AT&T
Lee BRATCHER – Head of Policy and Government Affairs, Cipher Digital
Sharmila BOKKA – Senior Electrical Engineer, Crusoe
MODERATOR: José RODRÍGUEZ – Chief Operating Officer, New Era Energy & Digital

AI workloads can change power requirements quickly. The panel examined unpredictable spikes from AI training clusters and the risk of stranded capacity. It also considered the problems that come with overbuilding or underbuilding infrastructure. As a result, the discussion focused on dynamic load management. It also stressed closer alignment between infrastructure planning and AI deployment cycles.

Day two

Time-to-Power: When the AI and Energy Dilemma Becomes a Geology Imperative

Fernando C. HERNANDEZ – Chairman of the Board, Society for Low Carbon Technologies

The second day opened with an unusual question: could geology help solve the time-to-power problem? The session pointed to several examples. Oracle has explored natural gas fuel cells, while Meta’s agreement with XGS Energy reflects growing interest in geothermal power. Deep Fission also entered the discussion. Meanwhile, Google aims to combine natural gas with carbon capture and storage to produce lower-carbon power for data centers.

Overcoming Delivery Bottlenecks Across the Data Center Ecosystem

Markus MUELLER – Field CISO, Nozomi Networks
Rajesh GOPINATH – Data Center Power Systems Engineering, Oracle
MODERATOR: Marc LALIBERTE – Principal, Special Projects, Continent 8 Technologies

Getting power available does not automatically mean a project can move ahead. This fireside chat looked at the bottlenecks that still slow data center delivery. It asked where utilities, developers, engineers, and EPCs can align earlier. It also considered whether EPC capacity has become as significant as power availability. Finally, the discussion looked at practical strategies that could help teams accelerate project delivery.

Accelerating Speed-to-Power: Standardizing AI Data Center Power Delivery Model

Scott YAPPEN – Senior Business Development Manager, Wärtsilä
Co-presenters: Sherif AbuSalem – Business Development Leader, Schneider Electric; Tom Mergen – Vice President and Growth Leader, Stanley Consultants

AI data centers can face grid interconnection delays of up to eight years. This session explored a standardised delivery model that combines power generation, electrical distribution and engineering. The model aims to improve speed-to-power, reliability and bankability. It also seeks stronger community acceptance. In turn, standardised partner delivery, pre-integrated equipment and repeatable execution could give developers a more consistent way to approach large AI projects.

Fireside Chat

Fernando C. HERNANDEZ – Chairman of the Board, Society for Low Carbon Technologies
MODERATOR: Marc LALIBERTE – Principal, Special Projects, Continent 8 Technologies

The programme included a fireside chat between Fernando C. Hernandez and Marc Laliberte. However, the event material does not provide a separate title or description for this discussion. Rather than add assumptions, the programme only identifies the format, participants, and their roles. The available material supports the speaker information, but it does not provide enough detail to describe the conversation itself.

Accelerating Time to Power: Sapphire Gas Solutions brings you natural gas-backed power and solutions without the wait

Jay SMITH – Executive Vice President of Business Development & Corporate Strategy, Sapphire Gas Solutions

Natural gas can offer another route when traditional grid access takes too long. This session introduced Sapphire Gas Solutions and its model for bringing natural gas to digital infrastructure projects, regardless of existing pipeline access. The approach supports projects from their initial power requirements through long-term expansion. In particular, the discussion showed how gas supply can form part of a wider strategy for getting power to digital infrastructure without waiting for conventional pipeline access.

Designing Ahead of the GPU Roadmap: Electrical Infrastructure That Scales With AI Load Density

Sharmila BOKKA – Senior Electrical Engineer, Crusoe

AI workloads do not behave like conventional data center loads. That difference can change almost every upstream power decision. This session looked at how poor load characterisation can lead to incorrect equipment specifications and costly change orders. It also examined battery storage as a load-driven design requirement. Beyond that, the discussion covered the challenges of combining on-site generation, storage and grid supply. The aim was to prevent today’s infrastructure from becoming tomorrow’s constraint.

The True Cost of Power: Pricing, Reliability Risk, and Market Design in Scarcity Conditions

Rajesh GOPINATH – Data Center Power Systems Engineering, Oracle
Uzair BAIG – Head of Power Strategy, Iron Mountain
Matthew FRANCIS – Construction Manager, Hoffman Construction Company
MODERATOR: Markus MUELLER – Field CISO, Nozomi Networks

Fast access to power can come at a price. This panel examined how capacity constraints affect PPA pricing and long-term power costs. It also looked at the choice between paying for speed and paying for certainty. Meanwhile, the discussion considered curtailment, congestion pricing and reliability events across different ISOs. These factors can influence site selection and investment decisions, especially when developers compare power options in scarce markets.

The organisations behind the event

The partner group reflected the range of systems that data centers need as they expand. Some focus on cybersecurity. Others work on transformers, power conversion, on-site generation, or energy management. Together, they cover several practical pieces of the power and infrastructure puzzle.

Nozomi Networks – Bronze Sponsor
Claroty – Bronze Sponsor
Wärtsilä – Bronze Sponsor
Dragos – Bronze Sponsor
Sapphire Gas Solutions – Bronze Sponsor
FR3 – Exhibiting Partner
Dynapower – Exhibiting Partner
Maddox – Exhibiting Partner
SMA – Exhibiting Partner
ComAp – Exhibiting Partner
Capstone Energy+ – Sponsor

To sum up

There is no single answer to the power challenges facing data centers. Grid access matters, but equipment, permitting, forecasting, cybersecurity, and onsite generation matter too. At the same time, AI continues to change both the scale and shape of power demand. The summit brought these issues together and explored them from several practical angles. The discussions also showed why developers need to consider power requirements much earlier in the project.