A timeline of key AI-market and legendary-investor developments — structured, dated, and sourced.
Running ~$22.7B, Laffont trimmed Nvidia and Meta and added application leaders (e.g., Netflix), shifting toward AI-driven, higher-certainty monetization.
Appaloosa nearly doubled Amazon into its top position on accelerating AWS, while trimming Nvidia and AMD and adding Micron — a rotation from compute toward where AI revenue lands.
Sources: 24/7 Wall St. — Tepper trimmed Nvidia/AMD, doubled an AI stock
In Q1, Duan made Nvidia his #3 position and opened Palantir, Synopsys, CrowdStrike and Snowflake while exiting Alibaba and CoreWeave — a value investor tilting toward AI.
Sources: 腾讯新闻 — 段永平 Q1 持仓
Two very different investors both hold Amazon — dual “picks-and-shovels + application” exposure that monetizes AI as cloud revenue — a rare consensus name.
The macro legend added AMZN and GOOGL for two straight quarters rather than Nvidia or Palantir — rotating toward where AI revenue shows up.
Hyperscaler custom accelerators (Trainium, TPU, Maia) are projected to rise from ~21% (2025) to ~28% (2026) of the market, eroding merchant GPU share over time.
Nvidia still holds an estimated 85–92% of AI accelerators; management points the 2027 AI-chip opportunity to at least $1T, with Blackwell’s rollout central to the thesis.
Beyond AI compute (CoreWeave, Cerebras), ARK invested heavily in nuclear (e.g., X-Energy), framing data-center power scarcity as the next bottleneck and opportunity.
Each quarter: how the legends moved + Consensus Score changes, in one email. Subscribers get first access when Pro tools launch. Free, unsubscribe anytime.