Gavin Baker, founder and Chief Investment Officer of Atreides Management, has built a reputation for positioning ahead of the broader market on artificial intelligence trades.
Baker’s core investment thesis centers on AI as a long-term supercycle, not a speculative bubble, driven by electricity, semiconductor wafers, and compute infrastructure.
He argues that real alpha in the AI trade does not lie in large language models or consumer-facing chatbots, but in the physical enablers underlying those technologies.
Baker’s so-called picks-and-shovels approach targets GPU interconnects, memory, inference chips, advanced process nodes, and power supply infrastructure as the true bottleneck assets of the AI era.
Atreides Management holds concentrated positions in companies such as Nvidia (NVDA), Cerebras, Astera Labs (ALAB), Micron (MU), Unity (U), and Positron, reflecting that physical-infrastructure conviction.
Baker’s early positioning in Nvidia and SK Hynix, before many institutional peers recognized the AI hardware trade, has drawn significant attention from investors and analysts alike.
His message to those trying to navigate the AI investment landscape is direct and unambiguous: “Follow the money.”
To manage broader market exposure, Baker employs QQQ puts as a hedge, allowing Atreides to maintain concentrated AI infrastructure bets while limiting downside from wider equity market moves.
Before founding Atreides Management in 2019, Baker spent nearly two decades at Fidelity Investments, from 1999 to 2017, most recently managing the Fidelity OTC Portfolio.
His long tenure at Fidelity, overseeing a large technology-focused portfolio, gave Baker an early and detailed understanding of semiconductor cycles and enterprise software transitions.
Baker’s decision to rotate out of software positions before the broader market recognized slowing growth in that sector is viewed by peers as a defining example of his forward-looking analytical framework.
The Atreides strategy reflects a broader institutional shift toward treating AI as a multi-year capital expenditure supercycle rather than a short-term thematic trade driven by sentiment.
As hyperscalers and governments continue deploying massive infrastructure budgets, Baker believes the physical constraints of AI scaling will keep hardware and power enablers at the center of the investment opportunity for years ahead.
