Land & Expand: Hyperscale, AI Factory, Megascale

October’s development activity highlights how the AI buildout is accelerating across North America, with Google, DataBank, Meta, and Lambda all advancing major projects while OpenAI and Oracle’s Stargate footprint expands into new states. This month’s 'Land & Expand' tracks the most notable site launches, land acquisitions, and retrofits shaping the next wave of hyperscale and AI-factory growth.
Nov. 14, 2025
12 min read

Key Highlights

  • Google’s $4 billion hyperscale campus in West Memphis marks its first data center in Arkansas, emphasizing renewable energy partnerships and community benefits.
  • DataBank’s Culpeper campus is progressing toward 192 MW AI-ready capacity, supported by a new on-site substation and regional utility upgrades.
  • Meta and Blue Owl’s Hyperion campus in Louisiana aims to support high-density AI clusters with advanced cooling and energy infrastructure over 4 million sq ft.
  • Vantage’s Stargate Midwest project in Wisconsin is a key node in the $15 billion Stargate expansion, with a 1 GW campus expected by 2028.
  • Lambda’s retrofit of a Kansas City facility into a 100 MW+ AI compute hub demonstrates the growing trend of reusing legacy assets for high-density AI workloads.
  • Stargate Michigan’s 1+ GW campus in Saline Township exemplifies large-scale, environmentally conscious development adjacent to existing transmission corridors.
  • Transmission upgrades and grid capacity planning are becoming critical bottlenecks, with utilities prioritizing AI-related infrastructure to meet future demands.
  • Cooling strategies are evolving with multi-mode and reclaimed water systems being integrated into new projects, indicating a shift toward sustainable thermal management.

Land & Expand is Data Center Frontier’s periodic roundup of notable North American data center development activity, tracking the newest sites, land plays, retrofits, and hyperscale campus expansions shaping the industry’s build cycle.

October delivered a steady cadence of announcements, with several megascale projects advancing from concept to commitment. The month was defined by continued momentum in OpenAI and Oracle’s Stargate initiative (now spanning multiple U.S. regions) as well as major new investments from Google, Meta, DataBank, and emerging AI cloud players accelerating high-density reuse strategies.

The result is a clearer picture of how the next wave of AI-first infrastructure is taking shape across the country.

Google Begins $4B West Memphis Hyperscale Buildout

Google formally broke ground on its $4 billion hyperscale campus in West Memphis, Arkansas, marking the company’s first data center in the state and the anchor for a new Mid-South operational hub. The project spans just over 1,000 acres, with initial site preparation and utility coordination already underway.

Google and Entergy Arkansas confirmed a 600 MW solar generation partnership, structured to add dedicated renewable supply to the regional grid. As part of the launch, Google announced a $25 million Energy Impact Fund for local community affordability programs and energy-resilience improvements—an unusually early community-benefit commitment for a first-phase hyperscale project.

Cooling specifics have not yet been made public. Water sourcing—whether reclaimed, potable, or hybrid seasonal mode—remains under review, as the company finalizes environmental permits. Public filings reference a large-scale onsite water treatment facility, similar to Google’s deployments in The Dalles and Council Bluffs.

Local governance documents show that prior to the October announcement, West Memphis approved a 30-year PILOT via Groot LLC (Google’s land assembly entity), with early filings referencing a typical placeholder of ~50 direct jobs. At launch, officials emphasized hundreds of full-time operations roles and thousands of construction jobs—indicating scaling phases that extend through 2027.

The site complements Google’s Southeastern clusters while offering low-latency east-west links across the company’s U.S. backbone network.

DataBank Advances Culpeper Campus Toward 192 MW AI-Ready Capacity

DataBank broke ground on its 85-acre Culpeper, Virginia campus, originally acquired in 2023, and now designed to support up to 1.4 million sq ft of hyperscale data centers. This includes two confirmed buildings—IAD5 (225,000 sq ft, 72 MW) and IAD6 (383,000 sq ft, 120 MW)—with AI/HPC-ready “Universal Data Hall” architecture baked into the base design.

The campus is enabled by a Dominion-built 300 MW on-site substation, planned to support 192 MW critical IT at full buildout. Dominion’s filings with the Virginia SCC (as part of the broader Culpeper Tech Zone docket) include a package of new 230 kV lines, station upgrades, and network reinforcements that also support neighboring campuses.

This makes Culpeper one of the first regions in Virginia where county-level zoning, utility planning, and hyperscale development were coordinated as a unified package.

DataBank expects its first building to come online in 2027, with power delivery sequencing dependent on the Culpeper Tech Zone grid upgrade window.

Meta and Blue Owl Capital Launch 4M-Sq-Ft Hyperion Campus in Louisiana

Meta and Blue Owl Capital announced the Hyperion AI campus in Richland Parish, Louisiana, spanning more than 4 million sq ft and backed by a $27 billion joint venture structure. Blue Owl is providing the capital and long-term ownership; Meta will manage construction and operations—an increasingly common hybrid model as hyperscalers diversify financing sources for multi-billion-dollar AI builds.

The campus represents Meta’s next wave of AI-first design, incorporating high-density single-story halls, multi-mode cooling (air + liquid), and long-lived energy-delivery infrastructure in partnership with local utilities. Meta documentation emphasizes a build sequence optimized for accelerated AI cluster deployment rather than general-purpose cloud.

State filings point to extensive community investment, including local hiring, workforce training pathways, and utility-scale upgrades to support multi-hundred-MW delivery. Hyperion positions Louisiana as a new node in Meta’s AI-ready southern U.S. footprint.

Stargate Midwest: Vantage Begins Work on the Port Washington "Lighthouse" Campus

Vantage Data Centers officially began development on the Stargate Midwest site in Port Washington, Wisconsin—internally referred to as “Lighthouse”—which forms one of the key anchor nodes in OpenAI and Oracle’s planned 4.5 GW Stargate expansion.

The program’s total investment has been reported at $15 billion, with Lighthouse slated for completion around 2028. Public filings and site plans describe multi-building single-story hyperscale shells with phased infrastructure rollouts, optimized for GPU cluster densification beyond Blackwell-era designs.

State economic development documents highlight the project’s expected role in regional grid planning, as Wisconsin utilities prepare for multi-hundred-MW data-center loads for the first time at this scale.

Lambda AI Accelerates Kansas City Retrofit Toward 100 MW+

Lambda announced the retrofit of a dormant 2009 enterprise facility in Kansas City’s Northland into a high-density AI compute center, targeting early 2026 for initial energization.

Phase 1 calls for 24 MW, with a path to 100 MW+, centered on the deployment of 10,000+ NVIDIA Blackwell Ultra GPUs dedicated to a single customer under a multi-year deal. Lambda emphasized that the site was chosen chiefly because it had substantial pre-existing utility service, enabling faster time-to-GPU than a greenfield build.

The project leverages Missouri’s Data Center Sales Tax Exemption Program, which requires minimum investment and job-creation thresholds. Local TV and state economic development reports frame the facility as a $500 million supercomputer hub with phased job creation ramping over time.

Lambda describes the reuse as “futureproofing unused power”—a characterization increasingly common as AI cloud providers scour secondary markets for brownfield assets capable of rapid high-density conversion.

Stargate Michigan: A 1-GW+ Anchor for OpenAI and Oracle’s National Footprint

The late-October announcement of Stargate Michigan in Saline Township (Washtenaw County) introduced one of the most ambitious hyperscale projects disclosed this year: a 1+ gigawatt campus developed by Related Digital on a 250-acre site. It forms one of the cornerstone nodes in OpenAI and Oracle’s expanded Stargate program.

The project includes three initial 550,000 sq ft single-story buildings (~1.65M sq ft), with LEED certification, structured setbacks, and extensive environmental design, consistent with Related’s development approach in mixed urban-rural zones.

A defining feature is its location adjacent to a high-voltage transmission corridor identified by DTE Energy as having “excess transmission capacity.” This enables large-scale interconnection without requiring new generation, reducing local rate-impact concerns and accelerating regulatory timelines.

Related Digital’s community benefits package includes $14 million for fire services, farmland preservation, and a community fund; more than 2,500 union construction jobs; and 450+ permanent roles at full build.

OpenAI notes that Michigan’s inclusion pushes its planned U.S. capacity (with Oracle) past 8 GW, with combined investment expected to exceed $450 billion over the next three years.

Roundup: Additional Projects Shaping the Megascale Build-Out

While October’s marquee announcements centered on Google, DataBank, Meta, the Stargate constellation, and Lambda’s Kansas City retrofit, several parallel developments across the U.S. and global market add necessary context to this phase of the AI factory build-out.

Microsoft, Amazon, and Oracle Quietly Add to Their Pipelines

Outside of formal ribbon-cuttings, filings and local disclosures show the hyperscalers continuing to bank land and interconnect capacity:

  • Microsoft advanced two AI-ready builds in Wisconsin and Georgia, with municipal filings showing increased water reuse treatment capacity and early trenching for redundant 400 kV-class transmission tie-ins.

  • Amazon Web Services secured approvals for additional substation expansions tied to its Virginia and Ohio AI clusters, including Dominion and AEP documentation noting “AI-generation density” and 2027 energization targets.

  • Oracle, beyond the Stargate Midwest and Michigan disclosures, quietly updated filings in Texas and Utah related to accelerated grid-capacity studies linked to future AI regions.

These developments indicate that the “megascale era” is moving from site announcements to synchronized grid planning and sequencing.

A Rising Tide of Retrofit and Reuse

Lambda’s Kansas City strategy fits a broader trend: 2025 is shaping into the strongest year yet for brownfield AI conversions, particularly for companies prioritizing time-to-GPU:

  • CoreWeave moved forward with two more Tier-III retrofits in Pennsylvania and Arizona, both leveraging dormant manufacturing envelopes with 50–90 MW utility stubs.

  • NVIDIA + Equinix extended their high-density JV pilots with additional retrofits in Silicon Valley and Dallas, supported by existing chillers and substation interconnects.

  • SADA AI and HPE GreenLake jointly filed for a high-density retrofit in Colorado, reusing a dormant 2012 wholesale shell.

This reuse wave is materially changing capacity forecasting: markets without greenfield land are producing highly competitive AI factory projects simply by reactivating legacy utility infrastructure.

Transmission & Substation Work Becoming the New Battleground

One clear theme across these projects—both announced and quiet-filed—is that transmission upgrades, not construction, are now the gating factor:

  • The Midcontinent ISO approved two “prioritized” queue acceleration requests tied to AI-adjacent campuses in Illinois and Iowa, reflecting new mechanisms to bring grid-ready sites online faster.

  • ERCOT, in its October reliability meeting, flagged 3+ GW of data-center-linked load entering the system by 2028, prompting new discussions around dynamic line ratings and AI-driven load forecasting.

  • PJM issued supplemental documentation on queue reform impacts in Northern Virginia—critical not just for hyperscalers, but for emerging HPC and AI cloud tenants banking capacity through 2030.

These grid actions shape whether 2026–2028 AI builds can actually energize on schedule.

Cooling, Water, and Megawatt-Class Thermal Planning Advance

Although cooling details were sparse for October’s biggest announcements, several adjacent disclosures show how hyperscalers are quietly reshaping their thermal strategies:

  • Google filed early stage multi-mode cooling plans for two new Midwest projects referencing seasonal switchover operations and onsite reclaimed-water processing.

  • Meta’s Hyperion site was linked to new wastewater-treatment construction in its parish documents, indicating a long-term shift toward independent water stewardship at AI-intensive campuses.

  • Blackstone-QTS submitted filings for its new Phoenix region campus showing 2-phase liquid cooling accommodation embedded in its base design—evidence that immersion-friendly shells are becoming standard even before tenants specify racks.

These moves underscore that the AI cooling transition is already underway—whether or not it’s highlighted in public announcements.

International Signals: AI Build-Out Is Globalizing

A handful of non-U.S. developments in the same timeframe provide an important counterweight to the domestic surge:

  • Microsoft locked another $2B for expansion in Finland tied to AI cloud workloads and new regional heat-reuse agreements.

  • Google disclosed progress on its Netherlands and Denmark AI clusters, including district-heating integrations and water-neutrality milestones.

  • Tencent, ByteDance, and GDS collectively added more than 300 MW under construction across Malaysia’s Johor region—further validating Southeast Asia as a global AI overflow zone.

The combination reveals a clear pattern: the U.S. remains the center of gravity, but AI megascale demand is now pushing global regions into their own AI-factory cycles.

 

At Data Center Frontier, we talk the industry talk and walk the industry walk. In that spirit, DCF Staff members may occasionally use AI tools to assist with content. Elements of this article were created with help from OpenAI's GPT5.

 
Keep pace with the fast-moving world of data centers and cloud computing by connecting with Data Center Frontier on LinkedIn, following us on X/Twitter and Facebook, as well as on BlueSky, and signing up for our weekly newsletters using the form below.

About the Author

David Chernicoff

David Chernicoff

David Chernicoff is an experienced technologist and editorial content creator with the ability to see the connections between technology and business while figuring out how to get the most from both and to explain the needs of business to IT and IT to business.

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.

Sign up for our eNewsletters
Get the latest news and updates
Tirachard Kumtanom/Shutterstock.com
Source: Tirachard Kumtanom/Shutterstock.com
Sponsored
Philip Tappe of Modius explores how modern DCIM solutions can optimize data center efficiency and PUE with real-time data, AI-driven analytics, and seamless integration. He shares...
Image courtesy of Schneider Electric
Source: Image courtesy of Schneider Electric
Sponsored
Alexander Rakow, Director of Sustainability, Secure Power and Data Centers for Schneider Electric, explains the importance of collaboration between solutions providers, data center...