This Week in Data Center News 10.13.25

The global data center industry is undergoing a massive transformation, driven primarily by the explosive growth of Artificial Intelligence (AI), creating urgent demands for revolutionary power solutions and infrastructure. Recent major industry developments underscore this shift: Brookfield Asset Management is investing up to $5 billion to deploy Bloom Energy's on-site fuel cells, providing a rapidly deployable, behind-the-meter alternative to the aging grid. This move, along with a strategic partnership between ABB Electrification and NVIDIA to develop next-generation, gigawatt-scale data centers using high-voltage DC architectures, signals a critical industry pivot toward more distributed and efficient power delivery systems to handle unprecedented power densities. These technological shifts are essential as global data center demand is projected to surge to approximately 220 GW by 2030.

Meanwhile, infrastructure and regulatory bodies are scrambling to keep pace with the power demands of AI. PJM Interconnection, for example, is pursuing emergency fast-track rule changes to accelerate data center grid connections, though this approach has generated controversy by potentially favoring natural gas over renewable energy projects. Simultaneously, hyperscalers like Google are cementing their long-term infrastructure commitments, with a $9 billion investment to expand cloud and AI infrastructure in South Carolina. These massive investments, coupled with innovations in backup power like Solidion Technology's PEAK Series UPS system , highlight both the enormous opportunity and the persistent power-related challenges facing developers as they build out the foundation for the AI-driven digital economy.

Brookfield Asset Management to spend an estimated $5 billion to deploy Bloom Energy’s fuel cells that provide data centers onsite power 

Bloom Energy has secured a major breakthrough in the AI data center power market through a strategic partnership with Brookfield Asset Management, which will invest up to $5 billion to deploy Bloom's fuel cell technology across global AI data centers. This deal addresses a critical infrastructure challenge facing the AI industry: the aging U.S. electric grid's inability to quickly provide the massive power capacity required for AI operations. Bloom's fuel cells offer a compelling "behind-the-meter" solution, providing on-site power that can be rapidly deployed without grid connections and can run on multiple fuel sources including natural gas, biogas, or hydrogen.

For data center developers, this partnership represents a significant shift toward distributed power generation as a viable alternative to traditional grid-tied infrastructure. The collaboration will establish "AI factories" globally, with the first European site expected before year-end, and builds on Bloom's existing deployments with major players like American Electric Power, Equinix, and Oracle. The urgency of this approach is underscored by the industry's massive power requirements—Nvidia and OpenAI's recent partnership alone targets 10 gigawatts of data center capacity, equivalent to New York City's peak summer consumption.

The market's enthusiastic response, with Bloom Energy shares jumping over 20%, reflects the recognition that off-grid power solutions may be essential for meeting AI's explosive growth while avoiding grid bottlenecks and protecting consumers from rising electricity costs. As Nvidia CEO Jensen Huang emphasized, self-generated data center power can move much faster than grid-based solutions, making this approach increasingly attractive for developers facing tight deployment timelines and capacity constraints.

ABB Electrification announces partnership with NVIDIA to develop next-generation data centers

ABB has announced a strategic collaboration with NVIDIA to develop next-generation AI data centers capable of operating at gigawatt scale. The partnership focuses on creating advanced power solutions to support NVIDIA's planned 800 VDC power architecture for 1-megawatt server racks, addressing the massive power requirements of future AI workloads. This collaboration combines ABB's expertise in direct current (DC) distribution and solid-state electronics with NVIDIA's AI infrastructure needs, targeting the development of medium voltage uninterruptible power supplies (UPS) and DC power distribution systems that can efficiently handle unprecedented power densities.

The timing of this partnership aligns with explosive growth projections for the data center industry, with global demand expected to surge from 80 GW in 2024 to approximately 220 GW by 2030, representing over $1 trillion in projected capital expenditure. AI workloads are anticipated to drive roughly 70% of this growth, creating an urgent need for revolutionary power distribution technologies. ABB's recent innovations, including the HiPerGuard solid-state MV UPS and SACE Infinitus solid-state circuit breaker, position the company as a key enabler of this transformation, with approximately 40% of their electrification R&D focused on next-generation data center technologies.

This collaboration represents a critical industry shift toward higher voltage DC architectures that can support the extreme power densities required for AI computing infrastructure. For data center developers, this partnership signals the emergence of standardized solutions for gigawatt-scale facilities, potentially reducing deployment complexity and improving operational efficiency. The focus on 800 VDC architecture specifically addresses the power delivery challenges that have become bottlenecks in scaling AI data centers, offering a pathway to more sustainable and economically viable large-scale AI infrastructure development.

PJM pursues new rule changes that could fast-track data center applications, meet surging demand & reshape grid navigation

PJM Interconnection, the grid operator serving 13 Northeast and Mid-Atlantic states, is implementing emergency measures to address unprecedented electricity demand from data centers through its Critical Issue Fast Path (CIFP) process. The region faces a projected 32-gigawatt increase in electricity demand by 2030—roughly 20% growth—while new power generation struggles to keep pace through PJM's current multi-year review backlog. In response, PJM has proposed an expedited interconnection queue allowing 10 large projects annually to connect within 10 months rather than years.

However, the proposed fast-track criteria—requiring projects to generate at least 500 megawatts and be completed within three years—effectively favor natural gas plants while screening out most renewable energy projects that rarely meet these size and timeline requirements. This approach has drawn criticism from clean energy advocates who argue it undermines the aggressive renewable energy goals of states in the PJM region, including New Jersey, Maryland, Virginia, Illinois, and Massachusetts. For data center developers, this represents both opportunity and constraint: while the expedited process could accelerate grid connections, the preference for gas generation may conflict with corporate sustainability commitments and state clean energy mandates that increasingly influence data center siting decisions.

The controversy highlights a critical tension in data center development—the urgent need for reliable grid capacity versus long-term decarbonization goals. PJM's December submission to FERC will determine whether these emergency measures become permanent policy, potentially shaping the energy landscape for data centers across the region for years to come.

Google to invest $9 billion in South Carolina data center expansion

Google is investing $9 billion in South Carolina from 2026 to 2027 to expand its cloud and AI infrastructure. The investment will fund the expansion of its existing Berkeley County data center campus and complete two new campuses in Dorchester County. This move aims to strengthen Google Cloud's South Carolina region, which is a key part of its global network of 42 cloud regions.

For data center developers, this significant investment signals robust growth and opportunity in the region. The expansion is directly tied to supporting the increasing demands of AI and cloud services, positioning South Carolina as a critical hub in the digital economy. The project also includes a workforce development component, with Google planning to train over 160 new electrical apprentices by 2030, indicating a long-term commitment to both infrastructure and local talent. This development underscores a major trend of hyperscalers expanding their footprint in strategic locations to support global AI transformation.

Solidon Technology Inc. announces new UPS battery system called “PEAK series”, specifically for AI data centers

Solidion Technology Inc., a small-cap battery company, has announced the development of the PEAK Series, a new Uninterruptible Power Supply (UPS) system specifically designed for AI data centers. This system utilizes the company's proprietary silicon-carbon anode technology in its 5500 battery cell. For data center developers, the key takeaways are the potential for a 30% space reduction compared to conventional systems and a longer lifespan, which could lower total ownership costs through reduced maintenance and streamlined assembly.

From a development perspective, the promise of a more compact and efficient backup power solution is compelling, especially given the rapid expansion and high power demands of AI infrastructure. However, it's crucial to note the context of this announcement. Solidion is a company with challenging financials, including a significant year-to-date stock decline and falling revenue. Furthermore, recent corporate restructuring and non-compliance with Nasdaq audit committee requirements introduce governance and stability concerns.

The PEAK Series is slated for commercial availability in the first quarter of 2026, with early testing currently underway with select partners. While the technology's claims are attractive for future data center projects, developers should approach with cautious optimism, weighing the potential benefits against the company's financial and operational risks until the product is commercially proven.

Tools & Solutions for Data Center Developers

Discover how we address critical challenges like power availability and project siting, and explore our range of available solutions. Book a demo with our dedicated team.LandGate provides tailored solutions for data center developers. 

You can also visit our library of data center resources.