CoreWeave secures multi-year Anthropic contract for AI workloads

CoreWeave, a publicly traded AI cloud infrastructure company, announced a multi-year agreement with Anthropic to run Claude AI model workloads in its data centers. The rollout will occur in phases, with the potential to expand over time, according to CoreWeave’s announcement.

Shares rose more than 12% on the news, trading around $102.73 at the time of reporting, according to Yahoo Finance coverage.

The deal comes amid CoreWeave’s recent financing round and strategic pivot. The company completed an $8.5 billion capital raise led by Meta Platforms, with the borrowing structured around deployed computing capacity rather than the company’s GPU hardware. In practice, the financing emphasizes predictable cash flows tied to the scale of compute capacity rather than the asset value of the hardware itself.

CoreWeave has long prioritized AI compute over crypto mining. The company pivoted away from mining and rebranded as an AI infrastructure provider in 2019, a move that positioned it to capitalize on growing demand for scalable AI workloads as the crypto industry faced cyclical pressures and rising energy costs.

Key takeaways

• The Anthropic deal is designed as a multi-year engagement with a phased deployment, opening the door to further expansion if demand grows.
• The $8.5 billion capital raise, led by Meta Platforms, is collateralized against deployed compute capacity, signaling a shift toward cash-flow-based valuation in AI infrastructure rather than hardware-backed lending common in crypto mining.
• CoreWeave’s pivot from crypto mining to AI infrastructure aligns with broader industry trends favoring AI compute markets in an environment of mounting mining headwinds.
• Bitcoin miners face sustained economic pressures, with a notable share reportedly unprofitable, which reinforces the appeal of directing energy and computing resources toward AI workloads.
• Analysts and market participants note that AI workloads—especially large-language-model inference and training—have become a more attractive revenue driver than traditional mining in recent years.

CoreWeave and Anthropic: a phased deployment for Claude workloads

In a statement, CoreWeave described the collaboration as a long-term, multi-year engagement aimed at supporting Anthropic’s Claude family of models. The plan is to roll out the compute capacity in stages, with the potential to scale as Claude’s demand grows and as the two companies refine capacity planning and efficiency. The arrangement underscores the ongoing shift in the AI ecosystem toward specialized cloud operators that can deliver cost-effective, scalable infrastructure for model development, training, and inference. By aligning with Anthropic, CoreWeave signals its intent to remain at the forefront of AI-accelerated compute, where the timing and cadence of deployments matter for both model developers and infrastructure providers.

CoreWeave has previously positioned itself as a bridge between AI research and production-grade compute, emphasizing the ability to deliver high-performance, scalable resources to a diverse set of AI workloads. The Anthropic partnership complements a strategy that seeks to monetize large-scale AI activity through predictable, capacity-driven revenue streams, rather than relying solely on hardware ownership or crypto-focused cycles. While the exact terms beyond the phased rollout were not disclosed, investors will be watching for indicators of expansion, such as additional model families integrated into Claude workloads or cross-service collaborations with other AI developers.

Financing anchored to compute capacity signals a strategic pivot

The capital raise tied to deployed compute capacity reflects a broader financial premise: the income stability of AI compute assets can be more predictable than hardware-backed collateral in volatile tech cycles. By stressing capacity-backed financing, CoreWeave and its backers aim to capture recurring revenue from ongoing Claude usage, rather than relying on the resale value or utilization of GPUs alone. The arrangement aligns with Meta Platforms’ broader investment in AI infrastructure, and it signals continued appetite among major tech sponsors for AI-oriented compute assets as a strategic asset class.

Industry observers have noted that such structures could become more common as AI workloads grow and require turnkey, scalable capacity that operators can commit to long term. For CoreWeave, the approach may enhance revenue visibility and help fund further expansion of its data-center footprint to meet rising demand from large-scale AI deployments.

Mining headwinds push AI compute demand higher

The broader crypto sector continues to wrestle with a challenging macro backdrop. Bitcoin mining remains capital- and energy-intensive, with rising energy costs squeezing margins as crypto asset prices fluctuate. CoinShares’ mining research has highlighted that as many as 20% of Bitcoin miners may be unprofitable under current conditions, underscoring the difficulty of sustaining traditional mining operations in today’s environment.

Market participants have observed a shift of some mining capacity toward AI processing and other high-value compute tasks, particularly when energy prices become more favorable for AI workloads. Market-maker Wintermute has underscored the need for miners to find yield opportunities for their assets, including deploying crypto into DeFi protocols to shore up revenues in tighter macro cycles. The sector’s stress intensified after the October 2025 market crash, when Bitcoin slid from a peak near $126,000 to the low-$60,000s before stabilizing in the $70,000s range. In this context, AI compute demand appears increasingly attractive as a more predictable cash-flow engine for data-center operators.

Analysts have framed this dynamic as a structural shift: AI compute needs—quantities of scalable, dependable processing capacity—are increasingly displacing traditional mining activity as the dominant driver of data-center utilization and profitability. As Ran Neuner noted in market commentary, “AI is willing to pay more for electricity,” a factor that complicates the economics of mining and tilts the balance toward AI-centric infrastructure solutions.

What investors should watch next

The Anthropic deal adds a new layer to CoreWeave’s earnings narrative, linking revenue growth to a major AI model developer’s deployment cadence and efficiency improvements. Investors will look for clear milestones on Claude workloads—such as rollout scale, latency benchmarks, and energy efficiency—and for confirmation that capacity expansion aligns with Anthropic’s model-usage patterns. At the same time, the sector-wide shift away from mining toward AI compute will continue to influence capital allocation, asset mix, and financing terms across AI-focused data-center operators.

For miners and AI infrastructure players alike, the key questions center on energy prices, the trajectory of AI compute demand, and the ability of data-center networks to scale while maintaining profitability. The CoreWeave-Anthropic alliance provides a concrete data point in a broader narrative: AI workloads may become the dominant driver of compute demand in the near term, with capital markets increasingly favoring capacity-backed models over hardware-centric financing in volatile cycles.

As the relationship between AI developers and compute providers deepens, observers will want to monitor how Anthropic’s Claude deployments scale in CoreWeave’s footprint, whether additional AI customers follow suit, and how this model of long-term, capacity-backed financing influences valuations and funding in the sector.

What remains uncertain is how broader regulatory and energy-market developments will shape the economics of AI compute versus crypto mining. Until then, CoreWeave’s latest collaboration with Anthropic serves as a tangible sign that AI-centric infrastructure—and the funding mechanisms that support it—are increasingly central to the next phase of digital technology deployment.