Yazılar

Nvidia invests $2 billion in CoreWeave to boost data center build-out

Nvidia has invested $2 billion in CoreWeave, becoming the AI infrastructure provider’s second-largest shareholder as the two companies deepen their partnership to expand data center capacity across the United States. The announcement pushed CoreWeave’s shares up 9% in premarket trading, highlighting investor confidence in the growing demand for AI-focused cloud infrastructure.

CoreWeave is part of a group of so-called neocloud companies that supply specialized hardware and computing capacity for artificial intelligence workloads. Demand for these services has surged as enterprises accelerate AI adoption. Nvidia’s new investment is expected to help CoreWeave speed up the acquisition of land and power needed to construct large-scale data centers, with the company targeting more than 5 gigawatts of AI data center capacity by 2030.

The investment was made at a purchase price of $87.20 per share, adding roughly 23 million shares and nearly doubling Nvidia’s stake in CoreWeave. Nvidia had previously held a 6.3% stake, making it the company’s third-largest shareholder. Despite scrutiny over Nvidia’s investments in AI firms, CoreWeave said the funds would be used for data center expansion, research and development, and workforce growth, rather than for purchasing Nvidia processors.

Once a cryptocurrency miner, CoreWeave has transformed its business to focus on leasing Nvidia GPUs to technology and AI companies. CoreWeave’s chief executive said the expanded collaboration reflects strong and growing demand for Nvidia’s computing platforms across the AI ecosystem.

CoreWeave Cuts Revenue Forecast After Data Center Delay, Shares Drop

CoreWeave (CRWV.O), a cloud infrastructure company backed by Nvidia, trimmed its annual revenue forecast on Monday after delays at a third-party data center partner disrupted operations, overshadowing strong quarterly results driven by soaring demand for AI computing services.

Shares fell more than 6% in extended trading, after Chief Financial Officer Nitin Agrawal forecast 2025 revenue between $5.05 billion and $5.15 billion, down from a previous estimate of $5.15 billion to $5.35 billion. Analysts had expected around $5.29 billion, according to data from LSEG.

CoreWeave said the customer impacted by the delay agreed to extend the contract’s expiration date, ensuring the total deal value remains intact, though the company did not name the client.

Despite the setback, the company posted a strong September quarter, with revenue more than doubling to $1.36 billion, beating Wall Street expectations of $1.29 billion.

CoreWeave has emerged as a major infrastructure provider for AI-driven workloads, securing high-profile contracts such as a $14 billion deal with Meta Platforms and a $6.5 billion partnership with OpenAI, both of which rely on its vast GPU-powered cloud network.

Once focused on Ethereum mining, CoreWeave has successfully repurposed its powerful GPU infrastructure to fuel the global AI cloud boom. However, its rapid growth has also exposed challenges — including rising chip prices, competition for computing capacity, and high expansion costs.

The company now expects to more than double capital spending next year, investing between $12 billion and $14 billion to meet surging demand.

CoreWeave shares have more than doubled since going public earlier this year at $40 per share, giving the firm a market capitalization above $50 billion, though its operating margin slipped to 16% in Q3 from 21% a year earlier.

Microsoft Signs $9.7 Billion Cloud Deal with IREN to Boost AI Computing Power

Microsoft has signed a $9.7 billion cloud computing deal with U.S.-based data center operator IREN to expand its artificial intelligence infrastructure and ease ongoing computing bottlenecks. The agreement, which includes access to Nvidia’s powerful GB300 chips, underscores the growing global demand for AI processing capacity.

Shares of IREN surged nearly 25% to a record high following the announcement before settling up around 10%. Dell Technologies, which will supply Nvidia’s advanced chips and related equipment to IREN, also gained about 1%. Under the five-year agreement, Microsoft will use roughly $5.8 billion worth of IREN’s computing hardware and infrastructure capacity.

The deal allows Microsoft to scale its AI operations without immediately building new data centers or acquiring additional power resources — key hurdles that have limited the company’s ability to meet soaring demand for applications like ChatGPT and Copilot. The approach also helps reduce heavy capital spending on rapidly depreciating hardware as newer processors enter the market.

IREN operates data centers across North America with a total capacity of 2,910 megawatts. The Nvidia chips will be deployed in phases through 2026 at the company’s 750-megawatt facility in Childress, Texas, which will include liquid-cooled centers providing 200 megawatts of new IT capacity.

The deal follows Microsoft’s recent $17.4 billion agreement with AI cloud provider Nebius and reflects the company’s strategy to leverage “neocloud” partners such as IREN and CoreWeave to expand capacity. IREN said Microsoft’s prepayment will help fund its $5.8 billion Dell contract, though the deal could be canceled if deadlines are missed.