meta platforms, inc.

overview

Meta Platforms, Inc. is a multinational technology conglomerate that owns and operates the world’s largest social networking platforms including Facebook, Instagram, WhatsApp, and Threads. Originally founded as Facebook in 2004 by Mark Zuckerberg and fellow Harvard students, the company rebranded to Meta in 2021 to reflect its strategic pivot toward building the metaverse and advancing artificial intelligence. Meta operates a massive global data center infrastructure to power its family of apps, serving over 3 billion daily active users worldwide. The company has pivoted heavily toward AI infrastructure investment, positioning itself as a leader in open-source AI development through its Llama model family and building unprecedented compute capacity to train next-generation AI systems.

business model

Meta’s business model is predominantly advertising-based, generating approximately 98% of revenue from targeted digital advertising across its family of apps (Facebook, Instagram, WhatsApp, Threads, Messenger). The company leverages its massive user base and sophisticated data analytics to deliver highly targeted ads to businesses of all sizes. Unlike cloud hyperscalers (AWS, Azure, Google Cloud), Meta builds data center infrastructure exclusively for internal use to power its social media platforms, AI model training, and content delivery. The company’s Reality Labs division develops virtual and augmented reality hardware (Quest headsets, Ray-Ban Meta smart glasses) and metaverse software, operating at significant losses while representing Meta’s bet on the future of computing. Meta’s open-source AI strategy with Llama models positions the company to shape the AI ecosystem without directly monetizing model access, ensuring it maintains access to cutting-edge AI capabilities without lock-in to competitors’ closed systems.

data center profile

global footprint

us portfolio (from database)

projects by state

specialization

primary focus: hyperscale, ai-ml, cloud

key differentiators:

• Open-source AI leadership with Llama model family - freely available models competing with closed alternatives

• Largest open-source AI training infrastructure in the world with 1.3+ million GPUs

• Vertical integration strategy - builds entire stack from custom silicon (MTIA) to applications

• Open Compute Project leadership - shares data center hardware designs openly to drive industry innovation

• Does not sell cloud services - all infrastructure supports internal products serving 3+ billion daily users

financial highlights

strategy

corporate strategy

Meta’s corporate strategy has undergone a dramatic transformation from social networking pioneer (2004-2020), to metaverse-focused company (2021-2022), to AI infrastructure leader (2023-present). The 2021 rebranding to Meta reflected ambitions to build the metaverse as the next computing platform, but Reality Labs’ persistent losses ($16+ billion annually) and tepid market reception led to strategic rebalancing. By 2023, Meta pivoted decisively toward artificial intelligence, announcing it had 'reoriented entirely toward AI infrastructure' with investments dwarfing metaverse spending. The company's AI strategy emphasizes open-source model development through Llama rather than closed APIs, believing this prevents lock-in to competitors while fostering ecosystem adoption. Meta's 'all the above' infrastructure approach combines self-built campuses, third-party leasing, AI-optimized designs, and innovative power solutions including on-site natural gas generation. The strategy leverages Meta's unique advantages: unparalleled social graph data for AI training, vertical integration from silicon to applications, and advertising-based business model that doesn't require monetizing AI model access. CEO Mark Zuckerberg frames Meta's massive capex investments ($60-72 billion in 2025) as essential to maintaining competitiveness in AI, even as Reality Labs continues operating at significant losses representing bets on future computing platforms.

growth strategy

Meta’s growth strategy centers on AI infrastructure buildout at unprecedented scale while maintaining its social media advertising dominance. The company is deploying approximately 1 GW of new compute capacity in 2025, bringing total to 1.3+ million GPUs by year-end - the largest AI training infrastructure globally. This massive investment aims to train increasingly sophisticated Llama models that compete with OpenAI’s GPT and Google’s Gemini while remaining open-source. Meta’s Louisiana Richland Parish facility ($10B investment, 2 GW capacity) exemplifies this strategy - the company’s largest data center ever, purpose-built for AI workloads. Unlike competitors building incrementally, Meta is making concentrated bets on massive facilities in regions with available power and favorable incentives. The company pre-leased more data center capacity in H2 2024 than any hyperscaler, particularly in Ohio, demonstrating aggressive pursuit of external capacity to supplement owned facilities. Meta’s growth strategy also emphasizes renewable energy portfolio expansion (11,700+ MW contracted, one of world’s largest corporate buyers) and nuclear partnerships (1.1 GW Constellation deal, geothermal projects) to secure reliable, carbon-free power. International expansion continues in Europe (Sweden, Denmark, Ireland) and Asia Pacific (Singapore) but at smaller scale than US buildout. The strategy accepts near-term profitability pressure from massive capex in exchange for long-term positioning as AI infrastructure leader capable of training cutting-edge models without dependence on cloud provider competitors.

power strategy

Meta pursues an ‘all the above’ power strategy combining renewable energy PPAs, nuclear partnerships, advanced geothermal, and innovative on-site generation to secure massive energy requirements while achieving ambitious sustainability goals. Unlike competitors focused primarily on one approach, Meta simultaneously pursues multiple pathways recognizing no single solution can meet AI infrastructure’s energy demands. The company has contracted 11,700+ MW of renewable energy, making it one of the world’s largest corporate buyers, with 89 of 128 projects operational by end of 2024 and target to add 9.8 GW to US grids by end of 2025.

renewable commitment: Meta has matched 100% of its global operations with clean and renewable energy since 2020, four years ahead of its original 2025 goal. The company has contracted over 11,700 MW of solar, wind, battery storage, and geothermal projects globally, representing one of the largest corporate renewable energy portfolios. In 2024, Meta signed PPAs across Indiana, Illinois, Arkansas, and Louisiana for more than 1.2 GW of contracted power. Notable agreements include RWE’s 274 MW County Run Solar (Illinois) and 100 MW Lafitte Solar (Louisiana), both commissioning late 2025. Meta purchased 760 MW of Renewable Energy Credits from Invenergy for projects in Ohio, Texas, New Mexico, and Arkansas connecting 2024-2027. The company’s renewable strategy emphasizes additionality - bringing new clean energy to grids rather than purchasing existing renewable credits. Meta works directly with utilities like Entergy in Louisiana to develop new renewable generation capacity, committing to support at least 1,500 MW of new renewable energy on the grid.

nuclear partnerships:

• Constellation Energy - 20-year agreement securing 1,121 MW of emissions-free nuclear energy from Clinton Clean Energy Center in Illinois beginning 2027. Partnership includes 30 MW expansion of Clinton’s clean energy output. Without Meta’s backing, plant faced premature closure. Preserves 1,100+ local jobs and $13.5 million annual tax revenue.

• Advanced geothermal partnership with Sage Geosystems - first deployment of next-generation geothermal power at scale east of Rocky Mountains, delivering up to 150 MW of new geothermal baseload power.

major commitments

partnerships

power providers

technology partners

NVIDIA (AI Hardware / GPUs) : Primary GPU supplier for AI infrastructure. Meta deploying 1.3+ million GPUs by end of 2025, primarily NVIDIA H100 Tensor Core GPUs. Meta built two 24,000-GPU clusters in late 2023, one using Infiniband and one using RoCE networking. Training Llama 4 on supercluster of 100,000+ H100 GPUs. Meta sought to acquire 350,000 H100 GPUs in 2024 at estimated cost of $10.5 billion.

AMD (AI Hardware / GPUs) : Meta’s Grand Teton platform now supports AMD Instinct MI300x GPUs alongside NVIDIA, providing supplier diversification for AI workloads.

TSMC (Semiconductor Manufacturing) : Fabricates Meta’s MTIA (Meta Training and Inference Accelerator) custom AI chips using 7nm process. MTIA designed to optimize Meta’s specific inference workloads and reduce dependence on external GPU suppliers.

Open Compute Project (OCP) (Open Hardware Standards) : Meta co-founded OCP in 2011 to share data center hardware designs openly. Contributed Grand Teton (GPU platform), Catalina (Blackwell-based rack), Open Rack v3, and Grand Canyon (storage) designs to OCP community, driving industry-wide innovation and cost reduction.

Dell Technologies (Server Hardware) : Server hardware supplier for Meta’s global data center infrastructure.

leadership

Mark Zuckerberg

Founder, Chairman, and Chief Executive Officer

Founded Facebook in February 2004 while attending Harvard University alongside Eduardo Saverin, Dustin Moskovitz, and Chris Hughes. Dropped out of Harvard to focus on Facebook’s growth. Built company from dorm room project to $1.79 trillion market cap. Led company through IPO (2012), major acquisitions (Instagram, WhatsApp, Oculus), rebranding to Meta (2021), and strategic pivot to AI infrastructure (2023-present). Controls ~58% of voting power through dual-class share structure despite owning ~13.5% of economic interest. One of world’s youngest self-made billionaires.

Visionary founder who transformed social networking and now betting company’s future on AI infrastructure and open-source AI strategy. Personally driving $60-72B annual AI capex decisions and Llama model development. Unique control through dual-class shares enables long-term strategic bets despite short-term profitability concerns.

Javier Olivan

Chief Operating Officer

Joined Facebook in 2007, 15+ years before becoming COO. Succeeded Sheryl Sandberg as COO in August 2022. Previously served as Chief Growth Officer, overseeing Facebook’s international expansion and growth from millions to billions of users. Responsible for operational efficiency across Meta’s infrastructure including data centers, networking, and platform reliability supporting 3+ billion daily active users.

Critical operational leader managing Meta’s infrastructure buildout and ensuring platform reliability during massive AI infrastructure expansion. Oversees teams responsible for data center operations and efficiency improvements.

Susan Li

Chief Financial Officer

Joined Facebook in 2008 as finance analyst, 14 years before becoming CFO. Succeeded David Wehner as CFO in November 2022. Deep institutional knowledge of Meta’s financial operations and strategic planning. Manages financial strategy during period of unprecedented capex investment ($60-72B annually) while maintaining profitability and shareholder returns.

Steers financial strategy during Meta’s historic AI infrastructure buildout. Balances massive capex investments against profitability goals and shareholder expectations. Communicates capital allocation decisions to investors during period of significant spending increases.

competitive position

Meta occupies a unique position in the data center industry: unlike AWS, Azure, and Google Cloud which sell cloud infrastructure services, Meta builds hyperscale data centers exclusively for internal use supporting its family of social media applications serving 3+ billion daily active users. This fundamental difference makes direct competitive comparison challenging - Meta competes with cloud hyperscalers for power, land, construction resources, and AI talent, but not for cloud services customers. In AI infrastructure investment, Meta ranks among the top spenders globally with $60-72B capex in 2025, comparable to Microsoft and trailing only by magnitude. Meta's open-source AI strategy with Llama differentiates from OpenAI's and Anthropic's closed models, positioning Meta as champion of open AI development. The company leads in several dimensions: largest open-source AI training infrastructure (1.3+ million GPUs), pioneering Open Compute Project hardware designs, industry-leading water efficiency (0.18 WUE vs 1.80 industry average), and massive renewable energy portfolio (11,700+ MW). Meta's vertical integration from custom silicon (MTIA) to applications provides cost and optimization advantages. However, Meta's business model concentration (98$10B Louisiana facility represents the industry’s largest single-site investment, demonstrating Meta’s willingness to make concentrated infrastructure bets. Meta faces competitive pressure from cloud hyperscalers attracting AI talent and from AI-native companies (OpenAI, Anthropic, Mistral) building competing model capabilities.

strengths

• Massive user base generating unparalleled training data - 3+ billion daily active users across family of apps

• Open-source AI leadership with Llama models - freely available, reducing barrier to adoption and ecosystem lock-in

• Largest open-source AI training infrastructure globally - 1.3+ million GPUs by end of 2025

• Vertical integration from custom silicon (MTIA) to applications - optimizes entire stack for specific workloads

• Open Compute Project leadership - shares innovations openly, driving industry cost reduction and adoption

opportunities

• AI infrastructure leadership positioning - investments create competitive moat in AI capabilities

• Open-source AI ecosystem building - Llama adoption creates favorable AI landscape versus closed competitors

• Advertising revenue growth from AI-enhanced targeting and ad products - leveraging AI capabilities for core business

• Developer ecosystem expansion - Llama adoption by businesses reduces competitive threats from proprietary AI

• Nuclear and advanced energy leadership - pioneering next-gen power creates competitive advantage and cost savings

threats

• Advertising market cyclicality - economic downturns severely impact 98% of revenue

• Competition from TikTok, YouTube, emerging social platforms - user attention fragmentation reduces engagement

• Regulatory forced divestiture - FTC case could force Instagram/WhatsApp spinoffs, dramatically reducing user base

• AI model competition - OpenAI, Anthropic, Google, Mistral releasing competitive or superior models

• Talent war intensification - AI researchers commanding premium compensation, bidding wars with competitors

projects

Showing top 20 of 29 projects

sources

• Meta Platforms Investor Relations

• Meta Platforms, Inc.

• Meta Platforms, Inc. Form 10-K for Fiscal Year Ended December 31, 2024

• U.S. Securities and Exchange Commission (2025-01-30)

• Meta Data Centers

• Meta Platforms, Inc.

• Meta Sustainability

• Meta Platforms, Inc.

• 2024 Sustainability Report

• Meta Platforms, Inc. (2024-08)

• The Llama 4 herd: The beginning of a new era of natively multimodal AI innovation

• Meta AI (2025-04)

• Building Meta’s GenAI Infrastructure

• Meta Engineering Blog (2024-03-12)

• Meta’s open AI hardware vision

• Meta Engineering Blog (2024-10-15)

• OCP Summit 2022: Open hardware for AI infrastructure

• Meta Engineering Blog (2022-10-18)

• Meta and Constellation Partner on Clean Energy Project

• Meta Newsroom (2025-06)