Anthropic enlists Micron to co-design HBM, DRAM, and SSD optimized for Claude inference at scale.

Anthropic and Micron Technology have announced a strategic agreement under which Micron will adopt Claude AI models as both a daily operational tool and an assistant to oversee parts of its infrastructure stack. The arrangement is notable for its unconventional structure: while typical supplier-buyer deals involve capital flowing from customers to suppliers, here Micron is essentially investing in one of its largest customers.

The logic underpinning the partnership is compelling. Anthropic operates some of the largest and most memory-intensive inference fleets in existence, and its operational telemetry on how HBM bandwidth, DRAM capacity, and SSD latency actually constrain real frontier-model serving is data Micron cannot generate internally. By collaborating closely, Micron gains access to this optimization data and can leverage Claude to process it into actionable improvements across its organization.

According to Tom Brown, Anthropic's co-founder and chief compute officer, the partnership addresses core scaling challenges: "Our compute strategy depends on getting every layer of the stack right, and memory and storage are central to how efficiently we can train and serve Claude. Partnering with Micron means we collaborate closely on optimizing these systems for our workloads and secure the supply we need. As demand for Claude grows, this is how we scale our compute for the long term."

Notably absent from both companies' announcements is any discussion of computational storage—an increasingly critical component in AI inference workloads. Nvidia has moved ahead in this space, announcing at CES 2026 its Inference Context Memory Storage Platform, which uses BlueField-4 DPUs to extend GPU KV cache into NVMe SSDs. This silence appears deliberate: Micron benefits considerably from selling premium HBM to the highest bidder, and computational storage solutions would undercut its most lucrative product lines. For Anthropic's part, maintaining optionality across suppliers—it has existing deals with AWS, Google, SpaceX, Broadcom, Microsoft, and CoreWeave, plus strategic commitments with Nvidia—means tying itself exclusively to one computational storage solution would be strategically limiting.

The agreement has been well-received by investors, with Micron's stock rising approximately 6% following the announcement. Neither company disclosed the financial terms of Micron's investment or the supply commitments involved. The deal follows a broader pattern in AI infrastructure, where major players forge strategic partnerships combining capital, compute, and supply guarantees—similar to Microsoft's arrangements with OpenAI or Nvidia's commitments with Anthropic's competitors.