AI Accelerates Everything
The idea
I believe AI and more efficient computing will crack problems that used to be too hard, speeding up progress across the whole economy. So this basket owns both sides of that bet: the chips, memory, chipmaking tools and power behind the computing, plus the companies using it for drug discovery, robotics and design. It is spread across six groups with steady names as ballast rather than a few concentrated picks, roughly half tilted to computing, with a small cash reserve set aside to add on pullbacks. This is meant to play out over years, not months.
What it owns
- Every AI breakthrough runs on chips made by a small number of companies, so their makers get paid regardless of which AI company wins. · Taiwan Semiconductor ManufacturingIt manufactures the most advanced chips for every major AI chip company, so nearly every leading AI chip passes through its factories., NVIDIAMost AI computing runs on its chips, so growing adoption of AI in complex applications drives demand for its processors., BroadcomIt designs custom AI chips for big tech companies seeking alternatives to single-supplier dependency, plus networking gear that links thousands of chips together., Advanced Micro DevicesIt is the main alternative to the market leader for AI chips, competing on cost and efficiency as a core theme of this take.28%
- Designing medicines used to take a decade of trial and error, and AI is starting to shorten that timeline significantly. · Eli LillyIt is a large, profitable drugmaker putting AI to work on finding new medicines while maintaining strong core drug sales., AlphabetIt owns DeepMind and Isomorphic Labs pursuing AI drug discovery, builds its own AI chips, and runs a major cloud business serving AI workloads., SchrödingerIt sells software that simulates how molecules behave, letting drug and materials companies test ideas on a computer before running experiments., Recursion PharmaceuticalsIt is a smaller company built around AI and automated labs for drug discovery, a pure bet that AI can solve hard problems in medicine.16%
- Cheaper, smarter AI lets machines do physical work in hospitals, warehouses and factories where labor costs are high. · Intuitive SurgicalIt makes the da Vinci surgical robot, the leading platform for robot-assisted surgery in operating rooms, and improved computing capabilities expand what these systems can do., SymboticIt builds AI-run robot systems that automate warehouses for large retailers, representing a smaller, higher-risk exposure to warehouse automation., Rockwell AutomationIt sells the controls and software that run factories, a steady way to own the trend toward automated manufacturing with AI., TeradyneIt tests chips before they ship and owns Universal Robots, which makes robot arms for factories, sitting on both automation and chip testing.15%
- AI needs huge amounts of memory and new factories to make chips, and only a few companies supply the equipment and components required. · ASML HoldingIt is the sole manufacturer of the machines required to produce the most advanced semiconductor chips, making it essential to every new ai chip factory built globally., Micron TechnologyAI chips need huge amounts of fast memory, and it is one of only three companies making the most advanced kind., Lam ResearchIt makes the equipment that etches and layers material onto chips, and memory makers rely on it especially heavily for advanced production.14%
- AI data centers use enormous amounts of electricity, and power equipment is one of the slowest things to scale up. · VertivAI data centers run hot and consume enormous power, and it sells the cooling and power equipment that keeps them operating., GE VernovaIt makes the gas turbines and grid equipment needed to bring new electricity to data centers for AI workloads., Constellation EnergyIt runs the biggest fleet of nuclear plants in the US, providing steady, around-the-clock power that tech companies prefer for data centers.11%
- Engineers use this software to design chips and products, and AI makes it far faster at problems that were too complex before. · SynopsysIt makes the software engineers use to design chips and is adding AI tools that make chip design work faster and more productive., Cadence Design SystemsIt provides chip design software and simulation tools for designing products and molecules, adding AI-powered capabilities to speed engineering work.8%
- Cash8%