According to @StockMKTNewz, AMD CEO Lisa Su told a UBS conference that commodities like memory have become tighter, highlighting constrained availability of memory components. Source: @StockMKTNewz. This is material for AI accelerators such as AMD Instinct MI300, which rely on HBM3 memory, making memory availability a gating factor for shipments and data center deployments. Sources: AMD Instinct MI300 product brief; AMD. AMD has disclosed reliance on third-party component suppliers, including memory, underscoring supply-chain sensitivity for fulfillment and margins when components tighten. Source: AMD 2023 Form 10-K. Industry context supports the comment, as DRAM leaders previously indicated HBM demand exceeded supply in 2024, pointing to persistent tightness into major AI buildouts. Sources: Micron 2024 earnings commentary; SK hynix 2024 earnings. For crypto-exposed AI compute ecosystems, the update is relevant because networks like Render (RNDR), Akash (AKT), and Bittensor (TAO) are built around GPU and compute availability that depend on memory supply. Sources: Render Foundation documentation; Akash Network documentation; Bittensor documentation.