Insights
Turning legendary investors’ disclosed moves into judgments you can use. Data-driven, sourced, updated quarterly.
Edge AI stocks in 2026: as the brain leaves the data center, who collects the toll?
Edge AI (running inference on the device itself) is a ~$30B market in 2026 heading toward $100B+ by 2033; on-device inference alone is ~$25–35B. Behind Apple Intelligence, Nvidia’s Jetson and Qualcomm’s Dragonwing, Arm’s architecture sits like a toll everywhere. We use the 13F consensus lens to separate the real picks-and-shovels from the concept-chasers as AI moves from cloud to edge.
Read →AI advertising stocks in 2026: AppLovin surges, Trade Desk stumbles — who turns AI into ad cash flow?
AppLovin’s revenue rose 59% and its AI ad engine drove a 27% single-day jump; Meta’s ad revenue climbed 33% to $55B with 8M advertisers using AI creative tools; Alphabet’s ads hit $77.3B. One “AI-monetizes-advertising” theme, yet it split into surgers and laggards. We use the 13F consensus lens to separate real cash flow from crowded trades.
Read →Sovereign AI: nations are building “national champions” — how do ordinary investors find the beneficiaries?
Nvidia’s “sovereign AI” revenue tripled to over $30B in FY2026; Abu Dhabi’s MGX closed a $49B fund and Saudi Arabia’s HUMAIN — backed by a ~$925B sovereign fund — signed for up to 600,000 Nvidia GPUs. When nations, not just companies, become buyers, the capex floor rises. We map this second-order theme’s beneficiary chain — and whether the legends’ 13Fs are betting on it too.
Read →Memory stocks fell into a bear market in a week: is the AI memory trade over, or is this the dip?
In the week of July 13–17, Micron, Samsung and SK Hynix all fell 20%+ below recent highs, SanDisk dropped ~35% from its June peak, and the freshly listed SK Hynix ADR sat just 3.4% above its IPO price. We break down the sell-off with data: what triggered it, whether fundamentals (contract prices) actually changed, where the smart money that piled into memory last quarter now stands, and the three signals to watch next.
Read →AI drug discovery in 2026: 15–20 AI-born drugs race into Phase III — how can ordinary investors participate?
For AI drug discovery, 2026 is the year of validation: 15–20 AI-designed or AI-discovered programs are expected to enter pivotal Phase III trials, Isomorphic Labs raised $2.1B, and Lilly and Novartis are betting big. We map the public-market entry points — from Tempus (an ARK add) to Recursion to the picks-and-shovels play, Nvidia — plus the risks unique to this lane.
Read →The custom AI chip wave: who is eating into Nvidia, and how big is the Broadcom opportunity?
Google TPU, Meta MTIA, Microsoft Maia, OpenAI Titan — hyperscaler custom silicon is everywhere. In 2026 ASIC shipment growth (~+45%) runs near triple that of GPUs, and Broadcom holds roughly 60% of design partnerships. Public data plus legendary-investor positioning show what remains of Nvidia’s moat and who profits on the custom chain.
Read →Mid-2026 institutional flows report: where is smart money moving in AI?
Five directions to read mid-2026 institutional money: Berkshire’s great pivot, hedge funds rotating into the memory chain, power as the new compute, sovereign funds’ $120B entry, and the starkest bull-bear split on record. All from public 13Fs and reporting, fully sourced.
Read →AI × stablecoins/RWA: where two megatrends cross — and why the legends mostly sit out
Stablecoins near $300B with $3.4T monthly volume; on-chain RWAs around $27.6B. As autonomous agents begin to pay and settle on their own, AI and crypto genuinely cross — in 2025, 40 cents of every crypto-VC dollar went to a company also building AI. Yet through the 13F consensus lens, the legends mostly hold back. That restraint is itself a lesson.
Read →AI’s power gap: why nuclear became 2026’s hardest second-order AI theme
Data centers are racing toward 12% of U.S. electricity (from ~4.4% at end-2023); solar/wind intermittency can’t meet AI’s baseload need, so nuclear is being repriced. Microsoft, Amazon, Google, and Meta have signed nuclear PPAs; uranium holds near $86/lb. We view this hardest second-order theme through a long-term, consensus lens.
Read →Agentic AI stocks: from Copilot to Agentforce, who’s actually monetizing?
The hottest enterprise-software story of 2026 is “AI agents” — give a goal, they execute across steps. But behind the hype, CFOs are tightening AI budgets. We map the listed exposures (Microsoft, Salesforce, Nvidia, Alphabet) and apply the consensus lens.
Read →Is Nvidia still a buy? Start with what 8 legendary investors actually did
The Nvidia bull/bear debate is endless. Skip the opinions, read the actions: Duan Yongping added sharply while Tepper and Coatue trimmed — the disagreement itself is the signal. We reconstruct the legends’ split on NVDA from public 13F filings.
Read →Which AI stocks have the strongest investor consensus? The leaderboard
Instead of one analyst’s call, look at eight legendary investors’ collective actions. The Compass Consensus Score turns their public 13F into a 0–100 score per AI stock, refreshed quarterly — a different lens on “consensus AI leaders.”
Read →What AI does Buffett own in 2026? Berkshire’s AI exposure, decoded
Buffett “avoids tech”? Berkshire’s real AI exposure hides in cloud and platforms. We decode his AI-related holdings from public 13F filings and compare them with seven other legends.
Read →Humanoid robot stocks: how to get public-market exposure (and who the legends back)
The leaders — Figure, Unitree, 1X — aren’t public. So how do you get physical-AI exposure? A map of listed plays: Nvidia, Tesla, UBTECH, Xiaomi, Hyundai, plus an ETF (KraneShares KOID) — and which of them the legends actually hold, via the Consensus Score.
Read →What is Physical AI (embodied intelligence)? Why Huang calls it a $40T market
AI is stepping out of the screen into the physical world: embodied intelligence and humanoid robots. Jensen Huang calls humanoid robots a ~$40T market and robotics Nvidia’s second growth curve after AI; SoftBank’s Son calls physical AI the birthplace of the next trillion-dollar company. We view this newest theme through the investor-consensus lens.
Read →Duan Yongping’s 2026 AI bets: a gradual shift from Apple toward AI
A value investor anchored in Apple and Berkshire, Duan Yongping tilted toward AI from 2026 — adding Nvidia sharply, opening AI-software positions, and holding PDD. We reconstruct his AI sleeve from disclosed actions.
Read →Cathie Wood / ARK’s 2026 AI bets: compute, cloud, and the next generation
Growth-style investor Cathie Wood bets on AI very differently from the value camp: next-generation compute and cloud infrastructure. High volatility often means high disagreement — the Consensus Score shows whether her names are consensus or a lone bet.
Read →AI’s most overlooked winners: power, nuclear, and second-order exposure
AI ends at electricity. Soaring compute drives data-center power demand, turning energy, nuclear, and the grid into second-order AI beneficiaries — a thread chip-chasers often miss. We examine it through a long-term lens and the legends’ disclosed positioning.
Read →How to “copy” a 13F: the right way to mirror the legends — and 3 traps
“Copy the legends’ homework” sounds easy but is full of traps: 45-day lag, longs-only disclosure, no cost basis. We explain what a 13F is, three common traps, and the right way — consensus over any single holding.
Read →Is AI a bubble? Don’t read opinions — read where the legends put real money
The “AI bubble” debate never settles. Instead of opinions, quantify the legends’ actual moves: the AI Conviction Index measures how strongly eight investors back AI this quarter, while the split on Nvidia shows the market is far from a single answer.
Read →Why David Tepper bets big on Amazon: the “AI monetization” thesis
While many chase chips, Appaloosa’s David Tepper made Amazon a top position while trimming Nvidia and AMD — a “bet on AI monetization, not the compute arms race” thesis. We decode it with disclosed actions and the Consensus Score.
Read →