−10.9%
since Sep 30
The idea
I believe AI is moving off the cloud and onto the devices we already own, as open models shrink enough to run fast, private, and cheap on local hardware. To express that, this basket spans the device makers building the boxes, the edge silicon enabling on-device inference, the memory and storage that hungry local models depend on, and a small sleeve in the open-model and fabrication ecosystem. It's concentrated by design, with no single name dominating, and I'm holding 10% in cash to add if prices dip.
What it owns
- Device makers and PC vendors benefit when enterprises and consumers refresh their hardware to support on-device AI inference rather than relying on cloud-based processing. · AppleIt has the largest installed base of devices with proprietary neural processors and has committed to running AI inference on iPhones and Macs., NvidiaIts consumer RTX cards are the standard hardware for running open-source AI models on home computers and gaming PCs., Dell TechnologiesIt manufactures the corporate laptops and workstations that get replaced during enterprise AI PC upgrade cycles., HP Inc.It is one of the world's largest PC makers, so an AI-driven refresh cycle in laptops flows directly to its unit volumes and pricing., Super Micro ComputerIt builds compact servers that businesses install on-premises to run AI locally rather than renting cloud infrastructure., Corsair GamingIt sells memory, power supplies, and cases for DIY PC builds aimed at running AI models at home, representing the consumer-enthusiast segment of the edge AI trend.33%
- Processors designed for edge AI inference, whether NPUs in phones and laptops, networking chips, or specialized vision processors, see volume growth as AI workloads move from data centers to devices. · QualcommIts Snapdragon processors include dedicated neural processing units that handle AI workloads in phones and Windows laptops., Advanced Micro DevicesIt ships Ryzen AI processors with integrated NPUs into AI-capable laptops, and also sells consumer GPUs for local model inference., BroadcomIt supplies the networking and connectivity chips that edge devices depend on, plus custom silicon for companies building proprietary AI processors., Arm HoldingsIts processor designs power most phones and a growing share of laptops, so increased AI device volumes drive higher royalty revenue., Lattice SemiconductorIt makes small, low-power programmable chips that handle AI inference in cameras, laptops, and industrial devices where dedicated processors are impractical., AmbarellaIts processors run computer vision AI inside cameras and vehicles, enabling inference to happen directly on the device.27%
- Local AI models require more memory and storage per device than previous generations, creating demand for higher-capacity DRAM, SSDs, and storage drives as AI adoption spreads. · Micron TechnologyIt manufactures the DRAM that powers local AI inference, and AI PCs are shipping with more memory per unit than prior generations., Western DigitalIt manufactures hard drives that store AI models and the data they process on local devices and home setups., SandiskIt produces the NAND flash that goes into SSDs shipped with every AI PC and phone, benefiting from higher storage capacity per device., Seagate TechnologyIts storage drives hold the growing volume of data that local AI systems create and consume at home and in edge server deployments.19%
- Open-source AI models enable capable AI to run on consumer hardware, and chip manufacturers across the ecosystem benefit from the resulting volume growth regardless of which vendor wins individual market segments. · Meta PlatformsIts Llama open-source models are a primary reason capable AI can run on consumer hardware, enabling the local-inference ecosystem., Taiwan SemiconductorIt manufactures nearly every processor in this take, from Apple to Qualcomm to AMD chips, benefiting from edge AI volume regardless of chip-designer competition.11%
- Cash10%