Data Center Insights: Phillip Marangella, EdgeConneX
The Data Center Frontier Executive Roundtable features insights from industry executives with lengthy experience in the data center industry.
Here’s a look at the Q4 2025 insights from Phillip Marangella, Chief Marketing and Product Officer, EdgeConneX.
Phillip Marangella is the Chief Marketing and Product Officer at EdgeConneX, where he develops and executes the company’s marketing and product strategies. With 25 years of experience in international marketing, strategy, and business development across the data center, telecom, and technology sectors, he brings significant expertise to his role. Phillip serves on the marketing advisory boards for Lokker, Salute, Data Center Coalition, and Infrastructure Masons. He holds a bachelor's degree in political science from the University of California, San Diego, and a Master’s in Multinational Commerce from Boston University.
Data Center Frontier: As liquid cooling shifts from pilot deployments to baseline design, what new coordination is required among equipment manufacturers, construction teams, and chemical/process experts to standardize safe, maintainable systems across global portfolios?
Phillip Marangella, EdgeConneX: The keyword is coordination:
- Coordination with the chip manufacturers on their latest reference designs and chip innovations.
- Coordination with customers and their deployment requirements and rack density requirements.
- Coordination with liquid to chip ecosystem vendors and partners on integrated liquid cooling solutions.
- Coordination with internal stakeholders on the design, commissioning, and operating of data centers leveraging liquid cooling in support of AI Factories.
And lastly: Coordination with our upstream supply chain to ensure long-lead items as identified in the design process are available when we need them (or securing acceptable alternative options when necessary.
Data Center Frontier: With hyperscale timelines collapsing and AI demand surging, how are owners and operators balancing speed-to-market with long-term cost of ownership, especially when power infrastructure and sustainability mandates move at different speeds?
Phillip Marangella, EdgeConneX: Speed of delivery and scale of capacity are the two key gating factors as we enter the Gigawatt-era in support of AI factories.
With the global data center market forecast to grow 6X and reach an annual take-up rate of 30GW, data center operators must develop new ways to deliver large volumes of capacity in much shorter timeframes.
That requires:
- New designs that optimize speed by using repeatable, standardized architectures across multiple sites;
- Foregoing traditional permanent structures, leveraging precast, prefabricated, and modular technologies;
- Making a significant up-front investment in critical supply chain equipment for power and data center infrastructure.
Data Center Frontier: As data centers are expected to evolve into energy hubs—producing, storing, and recycling power, water, and heat—what innovations or partnerships are emerging to make these facilities active participants in grid and resource stability?
Phillip Marangella, EdgeConneX: AI factories are being built at gigawatt scale. Markets like Texas and Ohio are forecast to add at least 5 Gigawatts of capacity next year. That represents as much as the entire global market only two years ago.
At that scale, it all starts with the power. If you don't have the power, then we can't turn up these AI Factories, and that means all the investment in these chips can't start generating tokens and revenue.
And while markets like Texas and Ohio have grid power, it's still not enough. Therefore, we need to augment grid power with alternative, on-site generation solutions. Whether it's natural gas, solar, wind, SMRs, geothermal, or energy storage solutions, they are all on the table and up for consideration to meet the massive power requirements.
For EdgeConneX, being supported by EQT and its other portfolio companies on the power and networking side means we can bring an integrated offering to market that can help simplify and accelerate the time-to-market of an AI-enabled data center solution.
Data Center Frontier: AI infrastructure now demands tight choreography among diverse disciplines, i.e. power, cooling, construction, chemistry, and digital systems. How are your teams aligning design and operations data across organizational silos to deliver performance and transparency from site prep to steady-state operation?
Phillip Marangella, EdgeConneX: We have essentially created an entirely new AI-enabled data center product that we call Ingenuity.
Keep in mind that for decades, we have been blowing cold air on servers. Little innovation was needed as rack densities remained in the single digits over that period.
With AI chips, rack densities are rapidly scaling into the triple digits, on the way to over 1 Megawatt per rack. In the not-too-distant future, we will see racks reach the same densities as those in some of our first edge data centers, built over a decade ago.
To respond to the pace and scope of change, we formed a large, cross-functional, internal AI task force nearly two years ago. The cross-functional team includes Engineering, Product, Operations, Commissioning, IT, Procurement, and others.
We have redesigned the data center. We have rewritten all of our operational procedures and created new procedures for areas not previously managed, like direct-to-chip liquid cooling.
We have established a new global training center to ensure the teams are prepared. We have updated and integrated our systems to monitor and manage our sites. We have documented the various operational demarcs between ourselves and our customers and subsequent SLA implications.
And the list goes on, all with the intent of ensuring we can continue to operate these AI factories.
About the Author
Matt Vincent
A B2B technology journalist and editor with more than two decades of experience, Matt Vincent is Editor in Chief of Data Center Frontier.



