On December 22, 2025, Alphabet announced a $4.75 billion acquisition that reveals more about the current state of AI development than any model benchmark or capability demonstration could: the company is buying Intersect Power, a data center and clean energy developer, for immediate cash plus assumption of debt. The deal includes multiple gigawatts of energy projects under development and a world-class team specializing in co-locating power generation with data centers. Most tellingly, it excludes Intersect’s existing operating assets in Texas and California, which will remain independent.
This is not primarily about acquiring technology or customers. It is about securing access to the resource that has become more critical than chips, talent, or algorithms in the AI race: electrical power. Google already held a minority stake in Intersect from a December 2024 funding round that also involved TPG Rise Climate and targeted $20 billion in renewable infrastructure investment by 2030. Now Alphabet is paying billions to acquire the development pipeline, the expertise, and—crucially—the ability to build power generation capacity “in lockstep with new data center load,” as CEO Sundar Pichai stated.
This deal exemplifies tech giants’ recognition that AI’s limiting factor is shifting from compute to power. No matter how sophisticated your models or how many GPUs you can manufacture, you cannot train or deploy AI at scale without reliable, abundant electricity. And the aging U.S. grid, struggling to meet flat demand for the first time in decades, cannot provide it fast enough.
The Energy Crisis Nobody Predicted
The Numbers That Changed Everything
Global data center electricity consumption reached approximately 415 terawatt-hours (TWh) in 2024, representing 1.5 percent of global electricity demand. The International Energy Agency (IEA) projects this will more than double to 945 TWh by 2030, reaching nearly 3 percent of global consumption. For context, that increase is equivalent to adding the entire current electricity consumption of Japan to global demand.
The United States faces particularly dramatic growth. U.S. data centers consumed 183 TWh in 2024, about 4.4 percent of total national electricity consumption. By 2030, this is projected to reach 426 TWh, representing 133 percent growth. The Department of Energy estimates data centers could consume between 6.7 to 12 percent of total U.S. electricity by 2028, depending on AI adoption rates.
This surge reversed decades of flat electricity demand growth. From the mid-2000s to early 2020s, U.S. electricity consumption remained relatively stable as efficiency gains offset growth. Starting in 2024, total annual U.S. electricity consumption hit a record high, with growth rates projected at 1.7 percent annually through 2026, driven substantially by data centers.
Why AI Consumes So Much Power
AI data centers differ fundamentally from traditional ones in their power requirements:
Computational Intensity
Training sophisticated AI models demands immense computing power. Training GPT-4 alone required around 30 megawatts of power continuously over the training period. OpenAI’s proposed Stargate initiative envisions data centers requiring 5 gigawatts—enough to power approximately 3.75 million homes.
A typical AI-focused hyperscale data center annually consumes as much electricity as 100,000 households. The larger ones currently under construction are expected to use 20 times as much, according to IEA projections.
Generative AI’s Energy Appetite
Generative AI consumes 10-30 times more energy than task-specific AI applications. An OpenAI ChatGPT query requires approximately 10 times as much electricity as a traditional Google search. When millions of users make billions of queries, that multiplier creates staggering aggregate demand.
In 2024, AI-specific servers in U.S. data centers used between 53 and 76 TWh of electricity—enough on the high end to power more than 7.2 million American homes for a year. If current adoption continues, AI applications could consume 35-50 percent of all data center power by 2030, up from 5-15 percent in recent years.
Power Density and Cooling
AI servers using high-performance GPUs generate substantially more heat than traditional servers, requiring advanced cooling solutions. Many AI data centers use liquid cooling rather than air cooling, which itself consumes additional energy. The power density—electricity consumption per square foot—in AI data centers far exceeds traditional facilities, creating infrastructure challenges that weren’t designed into existing buildings.
Regional Concentration and Grid Stress
Data center growth concentrates in specific regions rather than distributing evenly, intensifying local grid stress:
United States Dominance
China and the United States account for nearly 80 percent of global data center electricity consumption growth through 2030. U.S. consumption will increase by approximately 240 TWh (up 130 percent) from 2024 levels, while China will add around 175 TWh (up 170 percent).
By 2030, U.S. per-capita data center electricity consumption is projected to exceed 1,200 kWh per capita annually—roughly 10 percent of an average American household’s annual electricity consumption, and an order of magnitude higher than any other region globally.
Hotspot Markets
Northern Virginia hosts the largest concentration of data centers in the United States, creating particularly acute grid stress. The PJM electricity market stretching from Illinois to North Carolina saw data centers contribute an estimated $9.3 billion price increase in the 2025-26 capacity market. This resulted in average residential bills rising by $18 monthly in western Maryland and $16 monthly in Ohio.
One Carnegie Mellon University study estimates that data centers and cryptocurrency mining could lead to an 8 percent increase in average U.S. electricity bills by 2030, potentially exceeding 25 percent in the highest-demand markets like central and northern Virginia.
A 2024 Virginia legislature report estimated average residential ratepayers could pay an additional $37.50 monthly in data center energy costs. Harvard’s Electricity Law Initiative found that utility discounts given to tech giants for massive data centers can raise rates for other consumers, and if data centers fail to attract promised AI business, ratepayers could remain on the hook for subsidizing them.
The Grid Cannot Keep Up
The fundamental problem is that aging U.S. electrical infrastructure was not designed for this demand surge:
Connection Delays
Grid interconnection permits—approvals for connecting new power generation to the grid—face wait times stretching years in many regions. Tech companies are finding that securing power is often harder than securing land, equipment, or even regulatory approvals for data centers themselves.
Amazon’s attempt to buy a data center campus powered by Pennsylvania’s Susquehanna nuclear plant saw its grid interconnection permit twice rejected by the Federal Energy Regulatory Commission, illustrating regulatory friction even when power sources exist.
Transmission Bottlenecks
Even where generation capacity exists or can be built, transmission infrastructure often cannot deliver power where needed. Building new transmission lines faces NIMBY opposition, environmental reviews, permitting delays, and costs that make timelines stretch five to ten years for major projects.
Competing Demands
Data center power demand growth occurs alongside other electrification trends: electric vehicle adoption, heat pump installations replacing gas heating, and industrial onshoring. All these trends simultaneously stress grids designed for flat demand.
Alphabet’s Strategic Response: The Intersect Solution
What Alphabet Is Actually Buying
The $4.75 billion Intersect acquisition includes several critical components:
Development Pipeline, Not Operating Assets
Crucially, the deal excludes Intersect’s existing operating assets in Texas and its operating and in-development assets in California, which will operate as an independent company supported by TPG Rise Climate, Climate Adaptive Infrastructure, and Greenbelt Capital Partners. Alphabet is buying the future, not the present: multiple gigawatts of energy and data center projects in development or under construction.
This structure reflects sophisticated deal-making. Alphabet gains:
- Projects specifically designed for Google’s needs rather than generic assets
- Development expertise and execution capability
- Flexibility to customize future projects
- Avoidance of legacy commitments to other customers
The independent company retains valuable operating assets with established revenue streams, allowing those investors to capture ongoing cash flows while Alphabet focuses on building new capacity aligned with its requirements.
The Team and Expertise
Intersect’s “world-class team” led by founder and CEO Sheldon Kimber brings specialized expertise in co-locating industrial energy demand with power generation. Kimber has called Texas “the Disneyland of energy” for its abundant wind and solar resources, and his company has been marketing enormous data center sites in Texas to hyperscalers throughout 2024.
Intersect has $15 billion of energy assets in operation or under construction across the U.S. By 2028, Intersect projects representing about 10.8 gigawatts of power are expected to be online or in development—more than 20 times the electricity produced by the Hoover Dam.
Co-Location Model
The acquisition’s strategic value lies in Intersect’s co-location model: building power generation facilities directly alongside data centers rather than connecting to the general grid. This approach bypasses transmission bottlenecks, reduces interconnection delays, and provides dedicated power without competing for grid capacity.
As Sundar Pichai stated, Intersect will help Google “operate more nimbly in building new power generation in lockstep with new data center load, and reimagine energy solutions to drive US innovation and leadership.”
The Haskell County, Texas project exemplifies this model: a co-located data center campus and power generation facility where Intersect provides both infrastructure components. Google previously announced $40 billion in Texas investments through 2027, including new data center campuses in Haskell and Armstrong counties, and Intersect’s capabilities are integral to delivering power for that buildout.
Energy Source Strategy
Intersect focuses on what Alphabet describes as “multiple gigawatts of energy” from diverse sources:
Clean Energy Foundation
The December 2024 strategic partnership announcement with Google and TPG Rise Climate emphasized $20 billion investment targeting renewable power infrastructure by 2030. Intersect develops projects that co-locate industrial energy demand with natural gas and renewable generation, providing flexibility across different resource availability and regulatory environments.
Advanced Technology Exploration
Post-acquisition, Intersect will “explore a range of emerging technologies to increase and diversify energy supply,” according to Alphabet’s announcement. This includes:
- Advanced geothermal systems
- Long duration energy storage
- Gas with carbon capture and storage
- Grid connection acceleration using AI
- Energy efficiency and affordability programs in data center communities
Balancing Reliability and Sustainability
While Alphabet emphasizes clean energy commitments, the practical reality is that AI data centers require 24/7 reliable power, not intermittent renewables alone. As of 2024, natural gas supplied over 40 percent of electricity for U.S. data centers, renewables about 24 percent, nuclear around 20 percent, and coal around 15 percent.
The co-location model allows mixing sources: solar and wind when available, backed by natural gas or battery storage for reliability, potentially incorporating nuclear in future projects. This pragmatic approach prioritizes keeping data centers running over perfect environmental credentials.
The Broader Tech Response: Nuclear, Co-Location, and Billions
Alphabet’s Intersect acquisition fits within a broader pattern of tech giants scrambling to secure power through diverse strategies:
The Nuclear Renaissance
In 2024, tech companies made unprecedented commitments to nuclear power, both existing reactors and future small modular reactors (SMRs):
Microsoft Leads with Three Mile Island
In September 2024, Microsoft signed a 20-year agreement with Constellation Energy to restart Pennsylvania’s Three Mile Island Unit 1 reactor, dormant since 2019. The $1.6 billion project will generate over 800 megawatts of carbon-free power exclusively for Microsoft’s data centers starting in 2028.
The Nuclear Regulatory Commission is considering Constellation’s proposal for a three-year timeline to commercial operation. Unit 1 shut down in 2019 due to lack of economic viability, not technical issues, and had more than a decade remaining on its operating license. Constellation is already restoring equipment with 100 former employees on site, targeting over 700 positions by restart.
Amazon’s Aggressive Nuclear Bets
In October 2024, Amazon Web Services announced partnerships with Dominion Energy and X-Energy to develop and deploy 5 gigawatts of nuclear energy by 2040. Amazon invested $500 million in X-Energy’s 320-megawatt SMR project in Washington state and signed agreements supporting four projects with combined capacity of 320 megawatts in phase one, scalable to 960 megawatts.
Amazon also attempted to buy a data center campus powered by the Susquehanna nuclear plant in Pennsylvania in March 2024, though interconnection permits faced regulatory challenges.
Google’s SMR Agreements
In October 2024, Google announced collaboration with Kairos Power to build up to seven SMRs providing up to 500 megawatts of power. The first unit is expected online by 2030, with full completion by 2035. Google also committed early-stage capital to Elementl Power in May 2025 for three U.S. reactor sites totaling 1.8 gigawatts.
Oracle’s Gigawatt-Scale Plans
In September 2024, Oracle announced plans to construct a gigawatt-scale data center powered by three SMRs. Company founder Larry Ellison revealed that building permits had been secured and the project was in its design phase, targeting power generation exceeding 1,000 megawatts.
Meta’s Nuclear RFP
Meta issued a request for proposals in 2024 seeking 1-4 gigawatts of new nuclear capacity in the U.S. to support AI ambitions while meeting sustainability goals.
Why Nuclear?
Tech giants’ nuclear enthusiasm reflects several strategic considerations:
24/7 Reliability
Unlike solar and wind, nuclear provides constant baseload power regardless of weather or time of day. AI data centers cannot tolerate intermittent power, making nuclear’s reliability crucial.
Carbon-Free Credentials
Nuclear generates no direct carbon emissions, allowing tech companies to claim clean energy commitments while ensuring reliability that renewables alone cannot provide. For companies whose brands emphasize sustainability and environmental responsibility, nuclear offers a way to square the circle.
Long-Term Economics
While nuclear faces high upfront capital costs, operational costs are low and relatively stable over decades-long lifespans. For tech giants with long planning horizons and massive capital reserves, nuclear’s economics increasingly make sense, especially as natural gas prices fluctuate and carbon regulations tighten.
Small Modular Reactors Promise
New SMR designs offering 50-300 megawatts per unit provide scalability impossible with traditional gigawatt-scale reactors. Factory-built SMR components could halve construction timelines compared to conventional reactors. While SMRs face regulatory approval delays and remain largely unproven at commercial scale, they represent the nuclear industry’s attempt to solve traditional challenges around costs, timelines, and flexibility.
The Investment Scale
Combined AI infrastructure spending from Microsoft, Amazon, Alphabet, and Meta is expected to surpass $300 billion in 2025 alone. In 2024, major tech companies spent more than $200 billion on AI and cloud infrastructure.
Specific mega-projects include:
- OpenAI, SoftBank, and Oracle’s Stargate Project: $500 billion over four years for AI data center networks in Texas
- Microsoft: $80 billion in data center facilities in 2025
- Google: $40 billion in Texas through 2027
- Hundreds of billions more from Amazon, Meta, and other players
These investments dwarf historical technology infrastructure buildouts. The scale reflects both enormous projected AI revenues and genuine fear of falling behind in the AI race. Companies believe whoever secures power first will gain competitive advantages that determine market positions for decades.
The Hidden Costs: Who Pays for AI’s Energy Appetite?
Consumer Impact
While tech giants build private infrastructure, their power consumption affects everyone:
Rising Utility Bills
Data center power deals often involve utility companies providing discounted rates to attract massive industrial customers. Those discounts get subsidized by other ratepayers. Research from Harvard’s Electricity Law Initiative found that in some cases, if data centers fail to attract promised AI business or need less power than expected, ratepayers remain on the hook for subsidizing infrastructure built to serve them.
Nationally, electricity rates have already risen in recent years, partly from utilities replacing aging equipment for extreme weather resilience and cybersecurity, but data center growth adds additional pressure. The typical U.S. household was billed $142 per month in 2024, with projections showing continued increases.
Grid Reliability Concerns
As data centers claim more generation capacity and grid resources, questions arise about impacts on general grid reliability. If data centers get priority access during shortages, will residential and small business customers face rolling blackouts?
Federal Energy Regulatory Commission discussions about grid planning increasingly involve managing data center demands alongside traditional loads. Former FERC chairman Willie Phillips noted concerns about whether all projected data center demand is real, with some regions readjusting huge initial estimates.
Environmental Tradeoffs
Carbon Emissions Growth
Despite clean energy commitments, data center carbon emissions are rising. The IEA estimates data center emissions will climb from 180 million tonnes of CO2 currently to 300 million tonnes by 2035 unless cleaner energy sources are adopted faster than current trajectories.
Data centers represent one of few sectors where emissions are set to grow alongside road transport and aviation, even as most other sectors decarbonize. By 2030, data center emissions could reach 1 percent of global CO2 emissions in the IEA’s central scenario, or 1.4 percent in faster-growth scenarios.
Water Consumption
Data center cooling systems consume enormous water quantities. In 2023, U.S. data centers directly consumed about 17 billion gallons of water, with hyperscale and colocation facilities using 84 percent. Hyperscale data centers alone are expected to consume 16 to 33 billion gallons annually by 2028.
These figures exclude water consumed indirectly in electricity generation or semiconductor manufacturing. In drought-prone regions like the American Southwest, data center water demands create conflicts with agricultural, residential, and environmental needs.
Fossil Fuel Dependency
Natural gas remains the largest electricity source for data centers at over 40 percent, with coal still providing around 15 percent. While nuclear (20 percent) and renewables (24 percent) constitute significant shares, the buildout of new capacity often relies on natural gas for speed and reliability.
Elon Musk’s X supercomputer center near Memphis was found in April 2025 via satellite imagery to be using dozens of methane gas generators that the Southern Environmental Law Center alleged violated the Clean Air Act, exemplifying shortcuts some companies take when facing urgent power needs.
Uncertain Demand Projections
A fundamental question looms: Are tech giants’ power projections accurate, or are they overbuilding based on hype?
The Bubble Question
Grid Strategies estimates 120 gigawatts of additional U.S. electricity demand by 2030, including 60 gigawatts from data centers based on utility forecasts. But as GridUnity CEO Brian Fitzsimons noted, tech companies are “shopping the same big projects around to multiple utilities” as they seek quickest power access, potentially inflating apparent demand.
OpenAI CEO Sam Altman warned in August 2024 that the stock market faces an AI bubble with investors getting “overexcited.” If AI revenues fall short of projections, will data centers built to support them sit underutilized while ratepayers subsidize stranded infrastructure?
The Need for Firm Commitments
Grid Strategies president Rob Gramlich emphasized that utilities need firm financial commitments from data centers: “That is going to help us rationalize all these requests and get a better handle on the total estimate.” Without binding agreements, utilities plan based on uncertain projections that may not materialize.
Some regions have already readjusted projections downward after initial estimates. The uncertainty creates risks of both underbuilding (leaving companies unable to deploy AI) and overbuilding (wasting billions on unnecessary infrastructure).
Strategic Implications: The New Competitive Landscape
Power as Competitive Moat
Access to reliable, affordable power is becoming a defensive moat protecting AI market positions:
First-Mover Advantages
Companies securing power capacity first can build and operate data centers at scale while competitors wait years for grid connections. This timing advantage translates to earlier deployment of AI services, faster data collection for model training, and head starts in capturing market share.
Microsoft’s 20-year exclusive nuclear deal with Three Mile Island exemplifies this strategy: locking up dedicated capacity that rivals cannot access, ensuring power availability regardless of grid constraints.
Geographic Strategy
Power availability increasingly drives data center location decisions over traditional factors like proximity to customers or fiber connectivity. Texas attracts heavy investment because it offers abundant wind and solar resources, independent grid management, and land availability. Other states with aging grids or restrictive regulations lose opportunities.
This could reshape regional economic development patterns. States that facilitate power infrastructure gain tech investment, jobs, and tax revenue. States that don’t fall behind in the AI economy.
Vertical Integration
The Alphabet/Intersect deal represents vertical integration into power generation, following precedents from companies like Amazon (which owns solar and wind farms) and Microsoft (with its nuclear deals). This integration provides supply chain control and potentially lower long-term costs.
However, vertical integration into regulated utilities faces political and regulatory challenges. Electricity generation and distribution involve public interests that complicate pure private ownership models.
Market Concentration Risks
The enormous capital requirements for securing power at scale favor existing tech giants over startups and smaller players:
Barrier to Entry
A startup developing innovative AI cannot realistically negotiate nuclear power deals, buy data center companies, or invest billions in co-located generation. Cloud providers offer partial solutions, but reliance on Amazon, Google, or Microsoft infrastructure creates dependencies and limits differentiation.
This capital intensity could concentrate AI development among a handful of companies with balance sheets supporting multi-billion dollar infrastructure investments. The talent wars discussed in previous mega-deal analyses combine with power wars to doubly disadvantage challengers.
National Competitiveness
Countries that cannot provide abundant, affordable power for data centers risk falling behind in AI development. China, the United States, and Europe account for 85 percent of projected data center growth through 2030, partly reflecting power infrastructure capabilities.
Developing nations face particular challenges. Approximately 47 percent of academic institutions in high-income countries implemented AI-driven tools by 2023, whereas only 8 percent in low-income countries had done so. Power constraints are one factor perpetuating this digital divide.
Looking Forward: The 2030 Energy Landscape
Projected Growth Trajectories
If current trends continue, data centers will fundamentally reshape electricity markets:
U.S. Projections
BloombergNEF forecasts U.S. data center power demand will more than double by 2035, rising from almost 35 gigawatts in 2024 to 78 gigawatts. Average hourly electricity demand will nearly triple from 16 gigawatt-hours in 2024 to 49 gigawatt-hours by 2035. By 2035, data centers are projected to account for 8.6 percent of all U.S. electricity demand, more than double their 3.5 percent share in 2024.
Global Outlook
BloombergNEF expects global electricity demand from data centers to rise to 1,200 TWh by 2035 and 3,700 TWh by 2050. While these long-term projections carry inherent uncertainty, they’re grounded in fundamentals of increasing global digital adoption and data usage.
The balance of demand will shift geographically as digital adoption accelerates globally. While the U.S. remains the most important market currently, global data center demand is gaining ground.
Technology Wildcards
Several technology developments could significantly alter projections:
Energy Efficiency Breakthroughs
AI chip manufacturers are pursuing dramatic efficiency gains. If next-generation chips require half the power for equivalent computation, demand growth could slow significantly. Conversely, if efficiency gains enable even more ambitious AI applications, demand might accelerate rather than decelerate.
Liquid Cooling and Advanced Infrastructure
More efficient cooling technologies could reduce the energy overhead beyond just computation. Current estimates suggest cooling and infrastructure account for significant shares of data center power consumption beyond the servers themselves.
Algorithm Optimization
Improvements in AI algorithms that accomplish equivalent tasks with less computation could reduce power needs. Sparse models, mixture-of-experts architectures, and other innovations might deliver capabilities without proportional power scaling.
Quantum Computing
If quantum computers eventually handle certain AI workloads more efficiently than classical computers, power dynamics could shift dramatically. However, quantum computing remains far from practical deployment for most AI applications.
Policy and Regulatory Evolution
Government responses will shape how energy constraints affect AI development:
Grid Modernization
U.S. policymakers increasingly recognize that grid infrastructure requires massive investment to support electrification and data center growth. Federal infrastructure spending could accelerate transmission buildout, easing bottlenecks.
However, transmission projects face local opposition, permitting delays, and coordination challenges across jurisdictions. Even with political will and funding, physical infrastructure takes years to build.
Environmental Regulations
Stricter carbon pricing or emissions regulations would pressure data centers toward cleaner energy sources faster than current voluntary commitments. The EU’s carbon border adjustment mechanism and potential U.S. carbon pricing could change cost calculations favoring nuclear and renewables.
Conversely, the Trump administration’s emphasis on energy abundance and fossil fuel development could ease regulations limiting natural gas data center power, potentially slowing clean energy transitions.
Data Center Regulation
Some jurisdictions might implement data center-specific regulations addressing water consumption, power usage effectiveness requirements, or local grid impact assessments. Such regulations could slow buildouts in certain regions while favoring locations with supportive policies.
Conclusion: Power as the Prize
Alphabet’s $4.75 billion Intersect acquisition encapsulates the strategic reality of AI development in 2025: the limiting factor has shifted from algorithms and compute to electrical power. No matter how sophisticated your models or how many GPUs you can manufacture, you cannot train or deploy AI at scale without abundant, reliable electricity.
This deal represents pragmatic recognition that waiting for utility companies to expand capacity or hoping grid connections materialize fast enough carries unacceptable competitive risks. By acquiring Intersect’s development pipeline, expertise, and co-location capabilities, Alphabet gains ability to build power generation capacity “in lockstep with new data center load,” bypassing grid constraints that would otherwise throttle Google’s AI ambitions.
The broader pattern across tech giants—Microsoft’s nuclear deals, Amazon’s SMR investments, Oracle’s reactor plans, Meta’s nuclear RFPs—reveals an industry-wide scramble to secure power through any available means. These investments total hundreds of billions of dollars annually, exceeding historical infrastructure buildouts and reshaping energy markets.
The implications extend beyond tech companies’ competitive positioning. Consumer electricity bills are rising partially to subsidize data center growth. Grid reliability faces new stresses as concentrated industrial loads claim generation capacity. Environmental goals confront tensions between AI development and carbon reduction as companies embrace natural gas for reliability despite clean energy commitments.
The fundamental question is whether AI’s economic value will justify its energy appetite. If AI delivers transformative productivity gains, medical breakthroughs, scientific discoveries, and economic growth, the power consumption represents worthwhile investment in human progress. If AI falls short of hype, generating marginal improvements rather than revolutionary capabilities, we will have built massive energy infrastructure for uncertain returns while raising everyone’s utility bills.
Alphabet’s $4.75 billion bet on Intersect ultimately reflects confidence that AI’s value will justify its costs, including energy costs. The company is willing to vertically integrate into power generation, something tech companies historically avoided, because ensuring electricity access has become existential for AI leadership.
Whether that confidence is justified will become clear as the decade progresses and AI’s real-world impact on productivity, innovation, and economic value comes into focus. What’s already clear is that the AI race cannot be won without winning the power race first. Alphabet, Microsoft, Amazon, and others are spending billions to ensure they don’t learn that lesson too late.
Sources
- CNBC – Alphabet to acquire data center and energy infrastructure company Intersect (December 2025) https://www.cnbc.com/2025/12/22/alphabet-to-acquire-intersect.html
- Alphabet Investor Relations – Alphabet Announces Agreement to Acquire Intersect (December 2025) https://abc.xyz/investor/news/news-details/2025/Alphabet-Announces-Agreement-to-Acquire-Intersect-to-Advance-U-S–Energy-Innovation-2025-DVIuVDM9wW/default.aspx
- Axios – Google’s Alphabet buys data center, energy firm in $4.75 billion deal (December 2025) https://www.axios.com/2025/12/22/google-alphabet-data-center-intersect-power-deal
- Bloomberg – Alphabet to Buy Data Center Partner Intersect for $4.75 Billion (December 2025) https://www.bloomberg.com/news/articles/2025-12-22/alphabet-to-buy-data-center-partner-intersect-for-4-75-billion
- TechCrunch – Alphabet to buy Intersect Power to bypass energy grid bottlenecks (December 2025) https://techcrunch.com/2025/12/22/alphabet-to-buy-intersect-power-to-bypass-energy-grid-bottlenecks/
- Intersect – Alphabet Announces Agreement to Acquire Intersect (December 2025) https://www.intersect.com/news/alphabet-announces-agreement-to-acquire-intersect-to-advance-u-s-energy-innovation
- Yahoo Finance – Alphabet to buy clean energy developer Intersect in $4.75 billion deal (December 2025) https://finance.yahoo.com/news/alphabet-buy-data-center-infrastructure-161028901.html
- IEA – Energy demand from AI – Energy and AI Analysis https://www.iea.org/reports/energy-and-ai/energy-demand-from-ai
- Pew Research – What we know about energy use at U.S. data centers (October 2025) https://www.pewresearch.org/short-reads/2025/10/24/what-we-know-about-energy-use-at-us-data-centers-amid-the-ai-boom/
- BloombergNEF – Power for AI: Easier Said Than Built (July 2025) https://about.bnef.com/insights/commodities/power-for-ai-easier-said-than-built/
- MIT Technology Review – We did the math on AI’s energy footprint (September 2025) https://www.technologyreview.com/2025/05/20/1116327/ai-energy-usage-climate-footprint-big-tech/
- AI Magazine – AI Data Centres Will Drive a 165% Power Demand (September 2025) https://aimagazine.com/news/ai-data-centres-will-drive-a-165-power-demand-explained
- Carbon Brief – AI: Five charts that put data-centre energy use into context (September 2025) https://www.carbonbrief.org/ai-five-charts-that-put-data-centre-energy-use-and-emissions-into-context/
- Scientific American – AI Will Drive Doubling of Data Center Energy Demand by 2030 (April 2025) https://www.scientificamerican.com/article/ai-will-drive-doubling-of-data-center-energy-demand-by-2030/
- CNBC – Utilities grapple with multibillion question: How much AI data center power demand is real (October 2025) https://www.cnbc.com/2025/10/17/ai-data-center-openai-gas-nuclear-renewable-utility.html
- Department of Energy – DOE Releases New Report on U.S. Data Center Energy Use https://www.energy.gov/articles/doe-releases-new-report-evaluating-increase-electricity-demand-data-centers
- American Action Forum – AI Data Centers: Why Are They So Energy Hungry? (July 2025) https://www.americanactionforum.org/insight/ai-data-centers-why-are-they-so-energy-hungry/
- Data Center Frontier – How 2024 Re-Energized Nuclear Power (January 2025) https://www.datacenterfrontier.com/energy/article/55252205/how-2024-the-year-that-re-energized-nuclear-power-foretells-ongoing-new-nuclear-developments-for-data-centers-in-2025
- Nuclear Business Platform – Why Tech Giants Are Investing Billions in Nuclear Power (June 2025) https://www.nuclearbusiness-platform.com/media/insights/tech-giants-are-investing-billions-in-nuclear
- Fortune – Big Tech is the nuclear industry’s new best friend (December 2024) https://fortune.com/2024/11/21/tech-nuclear-energy-google-microsoft-amazon-ai/
- IEEE Spectrum – Big Tech Embraces Nuclear Power for AI and Data Centers (December 2024) https://spectrum.ieee.org/nuclear-powered-data-center
- Trellis – Amazon, Google, Meta and Microsoft go nuclear (June 2025) https://trellis.net/article/amazon-google-meta-and-microsoft-go-nuclear/
- Fierce Network – Energy is at the top of the data center agenda (December 2024) https://www.fierce-network.com/cloud/energy-top-data-center-agenda-heading-2025
- Data Center Frontier – Data Center Nuclear Power Update (November 2024) https://www.datacenterfrontier.com/energy/article/55239739/data-center-nuclear-power-update-microsoft-constellation-aws-talen-meta
- Modern Diplomacy – Nuclear Energy for AI Data Centers (February 2025) https://moderndiplomacy.eu/2025/02/17/nuclear-energy-for-ai-data-centers/
- Tom’s Hardware – Data centers to account for 9% of electricity demand in U.S by 2035 (September 2025) https://www.tomshardware.com/tech-industry/data-centers-to-account-for-9-percent-of-electricity-demand-in-the-u-s-by-2035-an-increase-of-5-percent-nuclear-power-could-help-sate-ai-demand
- Certrec – Why Are Tech Giants Investing in Nuclear Energy to Power Data Centers? (May 2025)https://www.certrec.com/blog/why-are-tech-giants-investing-in-nuclear-energy-to-power-data-centers/
