−10.4%
since Sep 30
The idea
I think AI is moving out of the cloud and onto the devices we already own, laptops, phones, home GPUs, as open models shrink and get good enough to run locally for privacy, speed, and cost. To express that, I own the device makers building the boxes, the edge silicon enabling on-device inference, the memory and storage that hungrier local models demand, and a smaller slice of who publishes open models and fabricates the chips.
What it owns
- These companies sell the phones, laptops, workstations, and servers that customers choose or upgrade to run AI inference locally rather than relying on cloud processing. · AppleIt controls a vast installed base of devices with proprietary neural silicon and has committed to running AI inference locally, so the take holds if Apple continues pushing compute onto iPhones and Macs., NvidiaIts RTX consumer graphics cards are the standard platform for running open AI models at home, so the take needs sustained local inference demand to drive consumer GPU sales., Dell TechnologiesIt sells corporate laptops and workstations that refresh when companies adopt AI-capable systems, so the take benefits if that upgrade cycle shows up in Dell's PC volumes and pricing., HP Inc.It is one of the world's largest PC makers, so an AI-driven upgrade cycle flows directly through its laptop volumes and selling prices., Super Micro ComputerIt builds compact servers that businesses deploy on-premises to run AI locally instead of using cloud capacity, so the take needs on-prem and edge AI server demand to grow., Corsair GamingIt sells memory, power supplies, and chassis for people building PCs to run models at home, making it the purest exposure to DIY AI compute becoming a real consumer category.33%
- These chip designers and manufacturers supply the processors, neural engines, and connectivity components that enable AI to run efficiently on devices at the edge rather than in distant data centers. · QualcommIts Snapdragon processors with dedicated AI engines are becoming standard in phones and Windows laptops, so the take needs Qualcomm to win sockets in the AI PC and premium phone markets., Advanced Micro DevicesIt ships Ryzen AI processors with built-in neural processing units into laptops and also sells consumer GPUs for local model inference, so the take depends on AMD gaining share in AI-capable PC designs., BroadcomIt supplies the networking and connectivity chips that edge devices depend on, plus custom silicon for companies building proprietary AI processors., Arm HoldingsIts chip designs power most phones and a growing share of laptops, so more AI devices running on Arm-based processors generates more licensing revenue., Lattice SemiconductorIt makes low-power programmable chips that run AI in cameras, laptops, and industrial systems where dedicated silicon fits better than a full processor., AmbarellaIts computer vision chips run inference inside cameras and automotive systems where processing happens on the device, so the take needs on-device AI in security and driver-assist systems to expand.27%
- Local AI models demand more memory and storage per device than prior generations, and this basket benefits if that capacity growth persists as AI devices scale. · Micron TechnologyIt manufactures the DRAM that AI-capable devices need in larger quantities, so the take depends on memory content per device staying elevated as AI PC adoption spreads., Western DigitalIt manufactures hard drives where model files and data for local AI live, so the take benefits if AI devices and home setups push storage capacity per unit higher., SandiskIt supplies the NAND flash inside SSDs that every AI PC and smartphone ships with, so the take needs flash demand and pricing to rise as AI devices carry larger and faster storage., Seagate TechnologyIts drives store the data that local AI generates and consumes, at home and in edge servers, so the take works if storage demand keeps rising with AI adoption.19%
- Open-source AI models and the manufacturing capacity behind them form the foundation that makes local inference practical and affordable for consumers and businesses. · Meta PlatformsIts Llama open models are the primary driver of capable AI running on consumer hardware, and broader adoption of open models accelerates the entire local inference premise., Taiwan SemiconductorIt manufactures nearly every processor in this basket, from Apple's to Qualcomm's to AMD's, so it wins on volume regardless of which designer prevails in edge AI.11%
- Cash10%