These are the AI hyperscalers fueling the demand—and building the compute capacity—that powers modern AI.

Key Industry-Wide Takeaways

• Record Capex Race: The top U.S. hyperscalers will collectively invest $300B–$392B in 2025 for AI/cloud infrastructure—by far the largest annual outlay in tech history.

• Compute Arms Race: AI hyperscalers require three times more energy than legacy data centers, driving deep partnerships with utilities, power producers, and private credit financiers.

• Hybrid Funding Models: Companies are increasingly turning to private credit, leasebacks, and joint ventures for flexibility and risk-sharing—Meta and Microsoft now mirror each other in $20B+ private capital packages.

• Regional Economic Impact: New data center clusters drive billions in economic development, job creation, and innovation, especially in “up-and-coming” markets.

• AI-First Scale: Proprietary silicon (TPU, Maia, Trainium, AMD ROCm), agentic AI infrastructure, and automation are critical for hyperscaler efficiency, cost, and technical leadership.

Amazon Web Services (AWS)

• 2025 Capex for AI/Data Center Infrastructure: Estimated at $100B+, with as much as $86B going to AI-specific cloud datacenters in the U.S. This is nearly a full year of AWS revenue reinvested.

• Major Projects: $30B in new data centers in Pennsylvania and North Carolina, including “AI innovation campuses” designed for large AI/LLM workloads.

• Technology Focus: Proprietary chips (Trainium, Inferentia) and AI-centric architectures.

• Growth Drivers: Surging cloud and AI demand; AWS AI business is now a multi-billion-dollar run rate, growing over 3× as fast as core AWS was at this stage.

Microsoft Azure

• 2025 Capex for AI Data Centers: Approximately $80B, with over half dedicated to U.S. installations.

• Key Initiatives: Rapid buildout of AI-centric data centers—adding over 1.5GW U.S. power capacity—supports Microsoft’s own AI apps and those of partners like OpenAI.

• Strategic Projects: $500B Stargate AI infrastructure joint venture (with OpenAI, Oracle, Softbank, MGX) aims for 10GW U.S. AI capacity through 2029.

• Technology: Custom Maia accelerators and growing investments in AI chips, large compute clusters.

Google Cloud Platform (GCP)

• 2025 Capex for AI/Data Centers: Raised to $85B in 2025 (up from earlier $75B plan), driven by exceptional AI and cloud demand.

• Key Projects: Accelerated construction of AI-optimized data centers; signature projects focus on advanced TPUs (e.g., Ironwood generation).

• AI Partnerships: Recently inked a major deal with OpenAI to provide additional compute for ChatGPT.

• Revenue Impact: GCP Q2 revenue grew 31% YoY, fueled by enterprise generative AI adoption.

Meta

• 2025 Capex (AI/Data Centers): $66B–$72B, nearly doubling previous year’s outlay. Focused entirely on AI infrastructure—largest investment in company history.

• Major Projects: “Titan” superclusters (Prometheus, Hyperion) in Ohio and Louisiana, targeting over 1GW and eventually up to 5GW of dedicated AI power.

• Financing: $29B hybrid equity/debt package with private credit partners to fund AI infrastructure, reflecting industry shift toward collaborative funding.

• Strategic Focus: Massive expansion to support superintelligence and generative AI for Facebook, Instagram, and future products.

Oracle Cloud (OCI)

• 2025 Capex & Commitments: Announced $25B plan for 2025–2027, primarily U.S. AI/data center buildout, including Abilene, TX, and Stargate project’s 4.5GW capacity build.

• Differentiators: Edge/AI clusters targeting regulated workloads (healthcare, telco, financial). Emphasis on cost-reducing RoCEv2 networking and modular expansion models.

• Strategic Partnerships: Leading GPU-provider for OpenAI’s Stargate initiative (alongside Softbank, MGX), targeting $30B annual revenue from cloud services by 2028.

CoreWeave

• 2025 Capex: Forecasting $20B–$23B in capital expenditures this year, including $6B+ for a new PA data center hub powering AI compute.

• Growth: 420% YoY revenue growth; backlog of $25.9B in performance obligations with major AI clients (OpenAI, IBM).

• Infrastructure: Scaling to 1.6GW contracted power; 33 AI data centers (28 in U.S.), aiming to double deployed power by end of 2025.

• Partnerships: Major deals signed with OpenAI ($11.2B+), IBM, and other hyperscalers.

OpenAI

• 2025 Key Investments: While OpenAI is not a traditional hyperscaler, it is a pivotal infrastructure consumer and investor through massive joint-venture deals.

• Strategic Projects: Stargate project—a joint venture with Softbank, Oracle, MGX—targets $500B investment by 2029. Involvement in gigawatt-scale capacity contracts (e.g., 4.5GW from Oracle) for next-generation model training.

• Dependence: Rapidly adding capacity from not just Microsoft but also Google Cloud and CoreWeave as compute demand outpaces Microsoft’s build rate.

Section Updated: August 10, 2025.

These firms bring AI hyperscale campuses to life through multi-billion dollar builds.

Key Industry-Wide Takeaways

• Surge in AI Hyperscale Construction: U.S. data center construction market is growing at a 15%+ CAGR, with AI and hyperscale investments accounting for over 50% of new project growth in 2024–2025.

• Workforce Impact: Each large-scale build (e.g., CoreWeave in PA, Meta in LA) brings thousands of skilled trades and professional jobs; AI construction is a major driver of U.S. workforce demand.

• Technology Transformation: AI is being used by leading builders to streamline scheduling, optimize site logistics, improve safety, and cut project delays and costs.

• Supply Chain/Power Constraints: Project delivery often runs up against utility bottlenecks and equipment procurement delays. Smart automation and pre-fabrication are increasingly necessary for competitive speed-to-market.

• Digital Twin Adoption: Jacobs and others are scaling up use of digital twins for data centers, maximizing operating efficiency and early issue detection.

Turner Construction

Flagship Projects:

• CoreWeave’s $6B AI Data Center Campus (Pennsylvania): Turner, in a joint venture with Wohlsen Construction, broke ground in 2025 on a purpose-built AI campus in Lancaster County. The 100MW facility (scalable to 300MW) is among the largest dedicated AI data centers in the Mid-Atlantic and will generate ~600 construction jobs.

• Meta’s $10B Louisiana Campus: Turner is also behind Meta’s record-breaking, clean-energy-backed AI center in Richland Parish, LA, slated as one of the largest hyperscale builds in the U.S..

• Vantage Data Centers (Ohio): Awarded a $2B expansion for high-density AI/ML-ready data halls.

• Meta (Indiana): Delivering an $800M hyperscale campus.

Innovation: Turner leverages AI-driven project management tools to cut project delays by 30% and deliver $50M in annual cost savings. 

DPR Construction

• AI Data Center Leader: Mortenson has delivered hyperscale campuses such as Meta’s 2.4M-sq-ft Eagle Mountain, Utah data center. The company is at the forefront of AI-fueled data center builds and reshaping power and cooling systems for high-density compute.

• Sector Insight: VP Maja Rosenquist highlights the race between data center build timelines and utility/power delivery, emphasizing selective and strategic development due to energy and labor bottlenecks.

• Design Trends: Mortenson’s leadership is driving rapid design adaptations for high-density GPU clusters and AI workloads.

Mortenson

• AI Data Center Leader: Mortenson has delivered hyperscale campuses such as Meta’s 2.4M-sq-ft Eagle Mountain, Utah data center. The company is at the forefront of AI-fueled data center builds and reshaping power and cooling systems for high-density compute.

• Sector Insight: VP Maja Rosenquist highlights the race between data center build timelines and utility/power delivery, emphasizing selective and strategic development due to energy and labor bottlenecks.

• Design Trends: Mortenson’s leadership is driving rapid design adaptations for high-density GPU clusters and AI workloads.

Hensel Phelps

• Confidential Hyperscale Campuses: Hensel Phelps is engaged in state-of-the-art, unnamed hyperscale data center campuses in Virginia, delivering secure, scalable facilities for Tier 1 cloud and AI clients. Due to NDAs, precise client names and investment figures are not public.

• AI in Operations: Through the Diverge Innovation Lab, Hensel Phelps collaborates with startups to pilot cutting-edge construction automation, robotics, and AI-driven project management.

Skanska

• AI Integration: Known for pioneering the use of AI and generative design in construction, Skanska pilots AI-powered site optimization (e.g., mass hauler fleet management), intelligent security, and BIM (Building Information Modeling) for precision and efficiency.

• AI Security: Uses AI-driven security on large civil and data center projects to deter threats, minimize downtime, and reduce costs.

Jacobs

• Digital Twins & NVIDIA Partnership: Jacobs is a leading force in AI hyperscale campus engineering, working with NVIDIA to deliver “AI factory” digital twin blueprints for U.S. data centers. The goal is to optimize power, cooling, and network integration for AI megacampuses.

• Booming Demand: CEO Bob Pragada reports record engagement and contract wins in advanced design-build services for AI and data centers, including confidential projects converting legacy sites into high-performance AI infrastructure.

Section Updated: August 10, 2025

Energy infrastructure partners serving massive AI compute demands.

Key Industry-Wide Takeaways

• Sector Capex Growth: U.S. utility capital expenditures are projected to rise by 15–20% year-over-year in 2025, largely due to AI and data center demand. Cumulative electric and gas utility capex for the largest 47 investor-owned utilities is forecast to surpass $1trillion between 2025–2029.

• Hyperscaler Demand: Massive tech companies (Amazon, Microsoft, Meta) are driving utility investments with multi-billion dollar data center and energy purchase agreements.

Entergy (LA)

• AI-Specific Capex: Public sources do not cite a standalone “AI hyperscaler” investment figure for Entergy, but industry analysts highlight a 15–20% capex increase across leading U.S. utilities in 2025, broadly tied to AI-driven energy and data center demand.

• Industry Position: Entergy is participating in grid modernization and new capacity additions, aligned with hyperscaler growth in key southern markets. However, no AI- or data center-specific major investment announcements by Entergy are reported as of August 2025.

NRG Energy

Strategic Initiatives: NRG Energy is actively expanding its footprint in the data center energy market.

• Menlo Digital Partnership: Recently announced a partnership with Menlo Digital to deliver up to 400MW of data center power capacity, with potential to exceed 1GW in future phases. This power will directly serve hyperscalers and AI data centers, particularly in high-demand markets (e.g., Northern Virginia, Silicon Valley, Dallas).

• Revenue Potential: Each 1MW of data center capacity typically generates $1–2million/year for energy providers. This partnership positions NRG for hundreds of millions in recurring, higher-margin revenue from hyperscale clients.

• AI & Grid Tech: NRG is deploying AI-powered grid management tools and growing its generation portfolio through acquisitions.

• Investments: Participated in a $2.5million investment in Equilibrium Energy, a platform leveraging AI to optimize energy portfolios.

Dominion Energy

Five-Year Capex Plan: Increased to $50.1billion for 2025–2029 (previously $43.2billion), with much of the growth directly attributed to surging data center (hyperscaler/AI) electricity demand.

• Data Center Demand: Data centers purchased 88% more capacity (now 19GW) as of December 2024, heavily driven by AI and hyperscale requirements.

• Virginia Focus: Over $40billion of the capex is allocated for Virginia to support data center and clean energy growth.

• Tech Collaboration: Entered a $500million joint venture with Amazon to explore small modular reactor (SMR) nuclear technology, adding potential new capacity for hyperscalers.

AEP / AEP Ohio

• AI/Data Center Direct Investment: No 2025 capex numbers specific to AI data centers found for AEP as of August 2025 in public sources; however, sector-wide capex increases apply.

• Industry Trends: AEP, like Entergy, is referenced among U.S. utilities ramping grid investments and deployment to address combined hyperscaler, AI, and advanced manufacturing needs.

• Regional Impact: Ohio is experiencing substantial data center expansions (ex: Aligned Data Centers), impacting AEP’s load forecasts and planning but without public, discrete numbers yet published.

Duke Energy (SC)

Capex Plan: Boosted its five-year investment plan to at least $83billion (previously $73billion), a 13.7% increase—with the main driver cited as data center and AI demand.

• More recent reports show plans to further increase infrastructure investments to $87billion, funded in part by selling non-core assets.

AI/Data Center Adaptation:

• Duke expects clients associated with data centers to account for up to 50% of its new connection pipeline by 2029, underscoring the AI-driven surge in load.

• Investment covers generation expansion (including new natural gas), grid modernization, and long-term clean energy resources for hyperscalers.

Financial Strategy: Includes $6.5billion of equity issuances through 2029 to support the capex acceleration.

Section Updated: August 10, 2025

These are the campus owners, builders, and critical facility specialists supporting AI deployments.

Key Industry-Wide Takeaways

• Unprecedented U.S. Hyperscale Buildout: Infrastructure companies rank as the main engines behind the trillion-dollar AI campus boom, with individual project investments often exceeding $2billion–$25billion per site or region.

• Power & Cooling Bottlenecks: Liquid cooling and advanced grid partnerships are critical to unlocking next-gen AI clusters, requiring new trusted partners like Schneider Electric and Vertiv to lead on power density and automation.

• Sustainability is Table Stakes: Major operators (Digital Realty, Vantage, STACK, Equinix) are focused on renewables, water-efficient cooling, modular builds, and digital twins to future-proof for AI and regulatory requirements.

• Strong Utility Collaboration: Strategic JVs between infrastructure companies and utilities (e.g., QTS/Blackstone, Compass/ComEd, Vantage/NV Energy) enable gigawatt-scale power guarantees.

• Gigawatt-Scale Growth: All players are scaling capacity in hundreds of megawatts—sometimes gigawatts—driven by AI demand from hyperscalers. Nearly every campus is built with liquid cooling, prefabricated modules, and high redundancy as standard.

QTS (Blackstone)

Investment & Expansion:

• Blackstone has deployed $25billion in Pennsylvania alone to build out QTS data center campuses and supporting energy infrastructure, with ambitions to catalyze $60billion in additional investment for AI capacity.

• QTS was transformed into a $60billion AI infrastructure powerhouse following Blackstone’s acquisition, rapidly scaling its development pipeline from $1billion to $25billion and commissioned capacity from 400MW to 3GW—becoming North America’s largest data center provider for AI.

• QTS is developing new sites in Iowa and Pennsylvania for next-gen AI deployments.

Digital Realty

Hyperscale AI Buildout:

• Digital Realty operates over 300 data centers globally, with rapidly expanding U.S. capacity. ~30% of new megawatt signings are directly tied to AI/hyperscaler demands.

• Projects emphasize sustainability, with future-proofing efforts for more than 300 facilities, renewable energy sourcing, advanced cooling, and low water usage for resource-intensive AI loads.

• Their operations span all leading U.S. hyperscale regions with continuous investment in scaling AI-specific halls and deployments.

Equinix (xScale)

Gigawatt-Scale U.S. Commitment:

• Equinix, via its xScale JV with GIC and Canadian Pension Plan Investments, is raising $15billion+ for hyperscale campus builds in the U.S., targeting 1.5GW of new AI-ready capacity for cloud and AI customers (AWS, Google, Meta, Microsoft).

• 16 xScale projects are underway worldwide, with new U.S. sites in Silicon Valley, Atlanta, and Dallas.

• Equinix emphasizes sustainability and efficient designs for high-density NVIDIA and AI workloads.

Vantage Data Centers

Recent Projects & Funding:

• $3billion investment in a new AI-optimized, 224MW Nevada campus (NV1), and $2billion+ in an Ohio campus (OH1) aimed at AI workloads, using modular designs with liquid cooling, low water use, and on-site substations for scalable GPU clusters.

• Secured $13billion in debt/equity recent funding to accelerate North American expansion, including multiple new AI/hyperscale sites in 2025.

• Industry leader in sustainable, scalable AI-specific builds (360–720W/sqft power densities).

STACK Infrastructure

Aggressive Hyperscale Expansion:

• $6billion+ in green financing so far in 2025, supporting flagship hyperscale campuses like the 1GW+ Stafford Technology Campus in Virginia (19 data centers across 500 acres).

• Announced a 36MW Chicago AI facility, designed for next-gen workloads—plus major footprint in Loudoun County, Portland, and Toronto, pushing total investment past $9billion in key markets.

• STACK projects focus on sustainability (renewables, low water use) and rapid, modular buildouts for U.S. hyperscale AI demand.

Compass Datacenters

High-Value Campus Builds:

• $10billion+ hyperscale campus investments in Chicago (former Sears HQ site) and Meridian, MS, hosting AI-specific deployments and supported by utility partnerships for power ramp-up.

• Model centers range from five to eight hyperscale buildings each designed to last 100+ years, supporting up to 500MW per campus.

• Focused solely on single-tenant, hyperscale AI campus delivery for leading tech firms.

NTT Global Data Centers / NTT DATA

AI Expansion & Market Strategies:

• NTT continues aggressive expansion with multiple new U.S. sites. Focus on integrating renewable energy, advanced cooling for high-density AI applications, and modular campus architectures.

• NTT projects align with demand from hyperscalers including AWS, Meta, Oracle, and other AI-focused clients.

Schneider Electric

Mission-Critical Power, Cooling & Automation:

• Launched new AI-ready data center solutions (EcoStruxure Pod and Rack): designed for liquid cooling and up to 1MW per rack—tailored for NVIDIA GB200 and high-density AI clusters.

• Powers rapid deployments with modular, prefabricated designs and full-scope automation. Noted partnerships include Compass Datacenters and NVIDIA for “digital twin” blueprints and scaled AI factory builds.

• Solutions address deployment speed, scale, and reduced onsite interventions (OPEX down 20%, interventions cut 40%).

Vertiv

AI Density Power & Cooling:

• Vertiv’s 360AI suite provides end-to-end power/cooling for up to 100kW per rack (with roadmaps for 800VDC power to support future NVIDIA and Intel platforms).

• Vertiv Unify enables campus-wide, cloud-connected management of power and cooling for AI environments.

• Prefabricated modules allow operators to scale from pilot AI clusters to full “AI factory” campus deployment, with deep partnerships with NVIDIA and Intel.

• Vertiv’s AI Solutions Roadshow and collaboration with industry leaders drive education and readiness for next-gen AI compute.

Section Updated: August 10, 2025