−8.9%
since Oct 10
The idea
I believe AI is moving out of the cloud and onto the devices we already own, so running models locally becomes the norm for privacy, speed, and cost. I'm expressing that across device makers, edge silicon, and memory and storage, since local inference needs both the chips and the RAM to back it up, plus a small slice in the companies publishing open models and fabricating the chips. I'm keeping 10% in cash to add on dips, avoiding oversized single-name bets, and treating this as a two-year view rather than something to hold forever.
What it owns
- Companies selling devices with on-device AI processors or enabling local model inference on consumer hardware. · AppleIt has the largest installed base of devices with dedicated neural processors and commits to running AI locally on iPhones and Macs., NvidiaIts RTX consumer GPUs are the default hardware for running open models at home and on local workstations., Dell TechnologiesIt sells corporate laptops and workstations that companies replace when upgrading for on-device AI capabilities., HP Inc.It is one of the largest PC sellers globally and benefits directly from an AI-driven device refresh cycle lifting volumes and pricing., Corsair GamingIt sells memory, power supplies, and cases for DIY PC builds used to run models at home., Super Micro ComputerIt builds compact servers that run AI locally on-premises instead of renting cloud capacity.33%
- Chipmakers designing processors and NPUs that handle AI inference at the device level rather than in data centers. · QualcommIts Snapdragon chips include dedicated NPUs that power AI processing in phones and Windows-on-Arm laptops., Advanced Micro DevicesIt ships Ryzen AI processors with built-in NPUs for AI laptops and makes consumer GPUs for local model inference., BroadcomIt provides networking chips edge devices depend on and custom silicon design for companies building proprietary AI processors., Arm HoldingsIts chip designs power most phones and a growing share of laptops, so edge AI adoption drives its royalty revenue., Lattice SemiconductorIt makes low-power programmable chips that handle AI tasks in cameras and small devices where full processors are unnecessary., AmbarellaIts chips enable computer vision AI inside cameras and vehicles, with inference happening on the device itself.27%
- Suppliers of DRAM, SSDs, and hard drives that local AI models require for speed and capacity. · Micron TechnologyIt makes DRAM that local AI models consume in large quantities, and AI PCs ship with materially more memory than prior generations., Western DigitalIt supplies hard drives that store model files and data on local devices and edge setups., Seagate TechnologyIt supplies hard drives that store data for local AI systems in homes and edge servers., SandiskIt makes NAND flash for SSDs in AI PCs and phones that carry larger, faster storage as standard.19%
- Companies enabling capable AI to run on consumer hardware through open-source models or foundational manufacturing. · Meta PlatformsIts Llama open models let capable AI run on consumer hardware people own, making local inference practical., Taiwan SemiconductorIt manufactures nearly every chip in this basket regardless of which designer wins, collecting volume regardless of competition.11%
- Cash10%