Apollo and Blackstone close $35 billion private credit package aimed at powering Anthropic’s next compute push

Apollo and Blackstone have reportedly closed a massive private credit facility valued at $35 billion to help fund Anthropic’s compute expansion, according to a market report published by Yahoo Finance’s investing outlet. The deal size indicates the scale of financing that large AI builders are seeking as they plan for additional servers, power, and accelerators to support growing model use and training needs.

The report frames the credit package as a way to fund compute capacity rather than an operating expense shift, meaning the capital is intended to enable new infrastructure buildout. In AI supply chains, that distinction matters, because compute expansion typically translates into additional demand for data center equipment, including specialized processors and high-bandwidth memory used to feed and serve large neural networks.

While the credit structure itself was not detailed in the posted report, the key economic idea is straightforward: private lenders are backing a major expansion effort by an AI company, and the resulting spending can flow to multiple layers of the hardware stack. For chip makers, that can appear as incremental opportunities tied to new systems, upgrades, and longer purchase horizons rather than a single quarter’s product cycle.

For Nvidia, the most direct linkage is its position as a leading supplier of GPUs and related software components used for training and large-scale inference. As Anthropic scales compute, demand for accelerator capacity is one of the channels through which the industry-wide spend can show up. Even if the credit funding is not explicitly earmarked for Nvidia hardware in public disclosures from the parties, the general market logic is that larger compute plans require more of the specialized chips that make those systems work.

Micron is implicated for different reasons, tied to memory used in AI servers. Large AI models are memory intensive, and data center memory bottlenecks can constrain how efficiently systems perform. If Anthropic’s compute buildout involves additional server deployments at meaningful scale, memory demand is one area that could see indirect pressure through the supply chain, depending on the mix of server configurations.

More broadly, the report highlights a sector pattern that has been growing since the surge in generative AI infrastructure spending: financiers are increasingly partnering with AI developers to underwrite capacity expansion when public markets are not the only path to capital. Private credit, in particular, can offer long-duration funding aligned with construction and equipment lead times, which are often longer than standard operating cycles.

Still, important specifics remain unclear from what has been published so far. The market report does not lay out the credit terms, maturity, investor commitments, or the operational schedule of Anthropic’s compute expansion. It also does not provide a granular mapping of how much of the resulting spending is expected to flow to any specific chip supplier, so any near-term read-through to individual companies should be treated as an inference rather than a quantified forecast.

Investors and industry watchers will likely focus on whether Anthropic’s compute plans translate into visible procurement activity across the hardware ecosystem, and whether major suppliers report changes in data center demand expectations tied to the timing of new infrastructure. In parallel, the financial community will watch how quickly the lenders deploy capital, since private credit deals of this size can have meaningful implications for funding conditions across technology-linked lending.