The thesis
We are in the early-to-mid stage of an AI supercycle. Not a 12-month narrative — a 10-to-15-year structural buildout. Every layer of the AI stack — compute, memory, networking, power, cooling, robotics, space — has bottlenecks. Each bottleneck creates a tradeable bloc.
But here is the part most miss: in this cycle, individual stock prices are dominated less by fundamentals and more by capital allocation flows from large funds routing through thematic ETFs. When a sector ETF takes inflows, all its constituents move together — whether their earnings deserve it or not.
QuantAbundancia maps that. Twelve bubbles, seventeen themes, and a growing universe of signals — each empirically validated.
How a bubble is validated
Anyone can name a sector. Few prove it actually trades as a bloc. Our method:
- Editorial taxonomy. We define each bubble by its capital-flow thesis (e.g. Datacenter Power = utilities and IPPs benefiting from AI-driven electricity demand).
- Daily price ingestion. Two years of OHLCV per ticker via Interactive Brokers.
- Pairwise correlations. 252-day Pearson correlation of daily returns for every pair.
- Hierarchical clustering. Within each bubble we cluster the names — if they don't actually move together, the bubble fails empirically.
- SPY-residualization. We strip market beta. What remains is idiosyncraticco-movement — real thematic flow, not just "all tech moves with QQQ."
Result: bubbles like Quantum (0.76) and Semi Equipment (0.82) score strongly even after stripping market beta. Bubbles like AI-Capex Hyperscalers mostly disappear under residualization — they're market beta wearing a thematic costume.
Three orthogonal axes
Bubbles
Empirically validated by 252-day co-movement. Trade as blocs.
Themes
Editorial groupings — narratives the market is telling itself. Some prove out empirically; others don't.
Watchlists
Cross-thematic conviction lists. The White House Basket. The Multibaggers. New ones added as the cycle evolves.
A single stock can sit in multiple bubbles, multiple themes, and multiple watchlists. The three axes are orthogonal — they capture different lenses on the same name.
Signals
The data engine writes one alert per rule per day when conditions fire. Today: Fibonacci-key proximity, new 252-day highs, new 252-day lows. Tomorrow: breakouts, volume spikes, correlation regime shifts, ETF flow surges.
The public alerts feed is the free tier. Telegram push and per-user subscriptions are coming as the paid tier.
Scope, today
- US-listed stocks and ETFs (ADRs included where IBKR resolves them)
- Daily close data, refreshed nightly after the US session
- ~135 tickers across 12 bubbles and 17 themes
- Pairwise correlations on 252 trading days
- Top-10 ETF holdings via yfinance + full holdings for iShares funds
- User-curated Fibonacci key levels for technical context
Crypto, forex, options, commodities are out of scope. Equity value-chain only — that's where the supercycle is buildable.
What we believe
Markets pretend to be efficient. They're not — they're crowded and structural. The largest moves of the next decade will be obvious in retrospect: where the capital had to flow because the buildout was non-optional. Power for AI. Memory for AI. Photonics for AI. The picks-and-shovels of a build with no ceiling in sight.
We don't predict prices. We map flows, validate them empirically, and surface the bottlenecks that are currently tradeable. The reader brings the conviction. We bring the rigor.
Roadmap
- Now — Live data, daily refresh, public bubbles/themes/watchlists, master heatmap, alerts feed.
- Next — Telegram push for alerts, per-user subscriptions, more rule types (breakouts, volume, correlation regime).
- Then — Subscription tier, mobile app, expanded universe (EU/Asia listings).
Get the daily digest.
A free email every morning: new alerts, biggest 252d highs, Fib-level approaches, and bubble shifts. No spam, no fluff. Telegram push and per-user signal subscriptions ship with the paid tier.
No spam. One email per day max. Telegram alerts coming with the paid tier.
QuantAbundancia is solo-built and self-hosted. No VC, no investors, no telemetry on your reading. Just a long bet on the AI buildout and the conviction that the rigor of a quant should be available to anyone who wants to surf the cycle.
Abundance, Quantified.