Meta Commits $35 Billion to CoreWeave and Reshapes the Global AI Race
On April 9, 2026, Meta announced the expansion of its partnership with CoreWeave in a deal that raised the total commitment between the two companies to approximately $35 billion — the largest cloud infrastructure contract for artificial intelligence ever signed between a big tech company and a specialized provider. The new agreement, valued at roughly $21 billion, covers AI computing capacity from 2027 through December 2032, adding to the previous $14.2 billion contract extending through 2031.
To put that number in perspective: $35 billion is more than the annual GDP of over 100 countries. It's the equivalent of buying 14 Maracanã stadiums made entirely of Nvidia chips. It's more money than NASA spent on the entire Apollo program adjusted for inflation.
What Happened
On April 9, 2026, Mark Zuckerberg and Meta's leadership publicly confirmed the expansion of their deal with CoreWeave, an American cloud infrastructure company specializing in Nvidia GPUs for artificial intelligence workloads. The new contract, valued at approximately $21 billion, secures Meta's access to CoreWeave's AI computing capacity from 2027 through December 2032.
This new agreement adds to the previous $14.2 billion contract, announced earlier in 2026, which covers the period through 2031. Combined, Meta's total commitment to CoreWeave reaches approximately $35 billion — a figure that exceeds the annual defense budget of countries like Colombia or Egypt.
The partnership is part of Meta's broader strategy to invest between $115 billion and $135 billion in capital expenditures (capex) in 2026 alone, nearly double what the company spent in 2025. This massive investment is primarily directed toward building and expanding gigawatt-scale data center campuses, acquiring AI chips, and developing infrastructure for what the company calls "agentic AI at global scale."
CoreWeave, for its part, operates data centers equipped with cutting-edge Nvidia GPUs — including H100 chips and the new Blackwell generation — offering on-demand computing capacity for companies that need to train and run AI models at massive scale. The company also maintains significant contracts with other AI giants, including Anthropic, the creator of Claude.
The Deal by the Numbers
| Component | Value | Period |
|---|---|---|
| Original contract | $14.2 billion | Through 2031 |
| New contract | ~$21 billion | 2027–2032 |
| Combined total | ~$35 billion | 2026–2032 |
| Meta capex 2026 | $115–135 billion | 2026 |
Context and Background
The partnership between Meta and CoreWeave didn't emerge from thin air. It's the result of a convergence of factors that have transformed the technology market over the past three years.
The Explosion in AI Computing Demand
Since the launch of ChatGPT in November 2022, global demand for GPUs and AI infrastructure has grown exponentially. Companies like Meta, Google, Microsoft, and Amazon entered an arms race for computing power, purchasing hundreds of thousands of Nvidia GPUs and building data centers at a frantic pace.
Meta, in particular, faced a unique challenge: with more than 3.5 billion monthly active users across its platforms (Facebook, Instagram, WhatsApp, and Messenger), any AI deployment needs to work at planetary scale. Training a single large language model can cost hundreds of millions of dollars in computing. Running that model for billions of users simultaneously multiplies that cost by orders of magnitude.
The Rise of CoreWeave
CoreWeave was founded in 2017 originally as a cryptocurrency mining company. When the crypto market slowed down, the company pivoted to offering its GPUs as cloud infrastructure for AI — a decision that proved visionary.
By 2024, CoreWeave was already valued at $19 billion. In 2025, after closing contracts with Microsoft, Nvidia, and Meta, the company completed an IPO that valued it at over $40 billion. Today, CoreWeave is considered the leading AI-specialized alternative to the three major cloud providers (AWS, Azure, and Google Cloud).
CoreWeave's differentiator lies in specialization: while AWS and Azure offer a broad range of cloud services, CoreWeave focuses exclusively on high-performance computing for AI, with Nvidia GPU clusters optimized for model training and inference.
Meta's MTIA Chip
Alongside the CoreWeave partnership, Meta is also developing its own AI chip, the MTIA (Meta Training and Inference Accelerator). The internally designed chip aims to reduce the company's dependence on Nvidia and its increasingly expensive and contested GPUs.
However, the MTIA is still in relatively early stages of large-scale deployment. While the proprietary chip matures, Meta needs guaranteed access to cutting-edge Nvidia GPUs — and that's exactly what CoreWeave provides. Meta's strategy is therefore twofold: invest heavily in external infrastructure (CoreWeave) while developing internal capacity (MTIA) for the long term.
The Race for Agentic AI
The term "agentic AI" has become the dominant buzzword in the tech sector in 2026. Unlike traditional chatbots that simply answer questions, AI agents are autonomous systems capable of executing complex tasks: making online purchases, scheduling appointments, managing business processes, negotiating contracts, and even making financial decisions on behalf of users.
Meta sees agentic AI as the next great leap for its platforms. Imagine an AI assistant on WhatsApp that can book a restaurant, compare product prices on Instagram Shopping, and organize your calendar — all without you needing to leave the app. For this to work for billions of people simultaneously, the required computational infrastructure is astronomical.
That's why Meta is willing to spend $35 billion with CoreWeave: it's not just about training AI models, but about running them in real time for a user base that represents nearly half the world's population.
Impact on the Population
The $35 billion deal between Meta and CoreWeave will have consequences that extend far beyond the corporate world. The effects will be felt by consumers, workers, investors, and even the environment.
Impact Table
| Aspect | Before | After | Impact |
|---|---|---|---|
| AI assistants | Basic chatbots with limited responses | Autonomous agents that execute complex tasks | Transformation in how billions of people interact with technology |
| Cloud market | Dominated by AWS, Azure, and Google Cloud | CoreWeave emerges as a fourth force specialized in AI | More competition and potentially more affordable prices for AI startups |
| Tech employment | High demand for traditional software engineers | Explosion in demand for AI and GPU infrastructure specialists | Professional reskilling needed for millions of workers |
| Energy consumption | Data centers already consume 2% of global electricity | Meta operates gigawatt-scale data center campuses | Pressure on power grids and acceleration of renewable energy demand |
| Privacy | Data used for targeted advertising | AI agents with deep access to personal data to execute tasks | New concerns about surveillance and sensitive data usage |
| AI startups | Limited access to expensive GPUs | CoreWeave offers specialized on-demand infrastructure | Partial democratization of access to cutting-edge AI computing |
For End Consumers
The most direct impact will be on the user experience across Meta's platforms. In the coming years, AI assistants integrated into WhatsApp, Instagram, and Facebook are expected to evolve from simple chatbots into truly autonomous agents. This means a user could ask their WhatsApp assistant to find the best price for a product, make the purchase, track the delivery, and even request a refund — all within a single conversation.
For small businesses that depend on Meta's platforms for sales and marketing, agentic AI could be transformative. Imagine an AI agent that automatically manages advertising campaigns, responds to customers, processes orders, and optimizes pricing — all without human intervention.
For the Job Market
The automation driven by agentic AI raises serious questions about the future of employment. Roles such as customer service, administrative assistance, social media management, and even basic financial analysis could be partially or fully automated by AI agents within the next five years.
On the other hand, building and maintaining the infrastructure needed for this revolution is creating thousands of new jobs in areas like data center engineering, GPU systems administration, AI engineering, and energy management for high-performance computing.
For the Environment
The gigawatt-scale data centers that Meta operates and plans to expand consume enormous amounts of electricity and water for cooling. A single gigawatt-scale data center campus consumes as much energy as a medium-sized city. With Meta planning to spend up to $135 billion in capex in 2026, the company's environmental footprint is under increasing scrutiny.
Meta claims to be committed to 100% renewable energy for its operations, but critics point out that the growing demand for electricity for AI is, in practice, delaying the energy transition by competing for renewable capacity that could be used to replace fossil fuels in other sectors.
What the Stakeholders Are Saying
Mark Zuckerberg, CEO of Meta
In an official statement, Zuckerberg declared: "We're building the infrastructure that will power the next era of computing. Agentic AI will transform how people connect, do business, and live their lives. Our investment with CoreWeave ensures we'll have the computational capacity needed to deliver this vision at global scale."
Michael Intrator, CEO of CoreWeave
Intrator celebrated the partnership expansion: "Meta is one of the most demanding and visionary clients in the world. This deal validates our thesis that the future of cloud computing is specialized, not generalist. We're building the infrastructure that will power the agentic AI revolution."
Jensen Huang, CEO of Nvidia
Although not directly involved in the deal, Jensen Huang commented on the broader trend during an investor conference: "What we're seeing is the beginning of a new industrial era. Companies like Meta and CoreWeave are building the factories of the future — intelligence factories. Demand for AI computing will grow for decades."
Market Analysts
Dan Ives, analyst at Wedbush Securities, called the deal "a defining moment in the AI race": "$35 billion is a number that would have been unthinkable two years ago. But when you consider that Meta has 3.5 billion users and wants to put an AI agent in the hands of each one of them, the investment makes strategic sense."
More skeptical analysts, like Aswath Damodaran of NYU, warned about the risks: "The history of technology is full of companies that spent fortunes on infrastructure for technologies that didn't materialize as expected. Meta is making a $35 billion bet that agentic AI will work at scale. If it doesn't, that money doesn't come back."
Regulators and Lawmakers
U.S. Senator Elizabeth Warren expressed concern about the concentration of power: "When a single company spends $35 billion to control AI infrastructure, we need to ask: who's overseeing this? Who ensures this technology will be used responsibly?"
In the European Union, regulators have already signaled that the deal will be analyzed under the AI Act, the European artificial intelligence legislation that came into force in 2025. The concern is that the concentration of AI infrastructure in the hands of a few companies could create barriers to competition and systemic risks.
Next Steps
Expected Timeline
The new $21 billion contract begins execution in 2027, with CoreWeave expanding its data centers to meet Meta's demand. The company has already announced plans to build new data centers in the United States, Europe, and Asia.
MTIA Development
Meta is expected to unveil the next generation of the MTIA chip in the second half of 2026. If the chip meets its performance targets, the company could gradually reduce its dependence on Nvidia GPUs and, consequently, on CoreWeave. However, analysts estimate this transition will take at least five years.
Agentic AI Rollout
Meta plans to launch the first full version of its AI agents for WhatsApp and Instagram by the end of 2026. The feature will initially be available in the United States and Europe, with global expansion planned for 2027.
Impact on CoreWeave's Stock
The $35 billion deal with Meta significantly strengthens CoreWeave's market position. Analysts expect the company, which completed its IPO in 2025, to see substantial stock appreciation as contracts with Meta and other major clients materialize into revenue.
Competitor Response
Google, Microsoft, and Amazon have already signaled they will increase their own AI infrastructure investments in response to Meta's move. The AI wars of 2026 are transforming into an infrastructure war, where computing capacity is as important as model quality.
Regulation
The U.S. Congress is expected to hold hearings on market concentration in AI infrastructure later in 2026. In Europe, the European Commission has already opened a preliminary investigation into the competitive impacts of long-term agreements between big tech companies and specialized cloud providers.
Closing Thoughts
The $35 billion deal between Meta and CoreWeave is more than a business transaction — it's a signal that the artificial intelligence race has entered a phase where investments are measured in tens of billions of dollars and planning horizons extend for decades.
Meta is betting that agentic AI will be as transformative as mobile internet was in the 2010s. If that bet is correct, the $35 billion invested in CoreWeave will be remembered as the foundation of a new technological era. If it's wrong, it will be one of the largest capital wastes in corporate history.
What's undeniable is that the world is changing. Gigawatt-scale data centers are being built at an unprecedented pace. AI chips are becoming the most contested strategic resource on the planet. And companies like Meta, Google, and Microsoft are redesigning global computing infrastructure for a future where AI agents will be as ubiquitous as smartphones.
For Meta's 3.5 billion users, the promise is of AI assistants that can do almost anything. For the planet, the question is whether we can power this revolution without compromising our environmental future. And for the tech market, the message is clear: in the AI race, those who don't invest billions will be left behind.
The deal also raises important questions about the future of work in the tech industry. As Meta and other companies invest billions in AI infrastructure, the implicit message is clear: the future belongs to fewer humans doing more with AI assistance. CoreWeave's own workforce of approximately 1,500 employees manages infrastructure worth tens of billions — a ratio that would have been unthinkable a decade ago. This efficiency-through-technology model is rapidly becoming the template for the entire industry, with profound implications for employment, skill development, and economic inequality.
For consumers, the Meta-CoreWeave partnership will likely manifest as increasingly sophisticated AI assistants embedded across Facebook, Instagram, WhatsApp, and Messenger. These agentic AI systems will be capable of booking travel, managing finances, negotiating deals, and even conducting complex research — all on behalf of the user. The question is whether consumers will embrace this level of AI autonomy or resist the erosion of human agency that it implies.
Sources and References
- Reuters — Meta expands CoreWeave deal to approximately $35 billion
- The Wall Street Journal — Meta's AI Infrastructure Spending Surges
- Bloomberg — CoreWeave's Growing Role in AI Cloud Computing
- CNBC — Meta plans $115-135 billion capex in 2026
- The Verge — Inside Meta's Agentic AI Strategy
- Financial Times — The AI Infrastructure Arms Race
