Boostio Research

Evidence first. Hypotheses labelled. Tested on paper.

Research tools for the AI build-out, fund disclosures and Boostio's own paper books, built only from public filings and market data. Every number carries its source and date, every idea says what would prove it wrong, and missing evidence is shown as missing.

Paper only: no broker, no real moneyNot investment adviceSources refreshed 2026-10-23 02:25 UTC
AI INFRASTRUCTURE DEMAND$601.5Bcash capex, 5 hyperscalers, 12 months+82.6% vs a year earlierPOWER DELIVERY · prices y/yGrid and data-center powerTransformers +7.8%Switchgear +12.3%Turbines +3.6%MEMORY · prices y/yHBM and DRAMChip producer prices −2.2%(all chip types, not memory only)NUCLEAR POWER · research onlyGenerators, fuel, reactorsVST · LEU · XE · OKLO · SMRGEVbacklog +37.0%GE Vernova (GEV): backlog +37.0%; research onlyPWRbacklog +75.1%Quanta Services (PWR): backlog +75.1%; research onlyETNrevenue +15.5%Eaton (ETN): revenue +15.5%; research onlyVRTrevenue +26.2%Vertiv (VRT): revenue +26.2%; research onlyMUrevenue +256.3%Micron Technology (MU): revenue +256.3%; research onlyAI INFRASTRUCTURE DEMAND$601.5Bcash capex, 5 hyperscalers, 12 months+82.6% vs a year earlierPOWER DELIVERY · prices y/yGrid and data-center powerTransformers +7.8%Switchgear +12.3%Turbines +3.6%MEMORY · prices y/yHBM and DRAMChip producer prices −2.2%(all chip types, not memory only)NUCLEAR POWER · research onlyGenerators, fuel, reactorsVST · LEU · XE · OKLO · SMRPOWER DELIVERY CANDIDATESGEVbacklog +37.0%GE Vernova (GEV): backlog +37.0%; research onlyPWRbacklog +75.1%Quanta Services (PWR): backlog +75.1%; research onlyETNrevenue +15.5%Eaton (ETN): revenue +15.5%; research onlyVRTrevenue +26.2%Vertiv (VRT): revenue +26.2%; research onlyMEMORY CANDIDATESMUrevenue +256.3%Micron Technology (MU): revenue +256.3%; research only
The lines are Boostio's hypothesis chain, not measured money flows. Numbers: SEC filings and BLS producer prices, with dates and links below. A green outline marks a paper-tested candidate; an orange one, a triggered invalidation rule. The dashed nuclear branch is research only: no paper trades.
Tools

Six more ways into the same evidence

AI Bottleneck Radar

Where the AI build-out could run short

Hyperscaler capital spending from SEC filings, producer prices for transformers, switchgear, turbines and chips, and the companies that might benefit in power delivery, memory and nuclear power, each with its uncertainties and what would invalidate it.

Combined capex +82.6% over 12 months · 5 candidates · 5 nuclear names, research only · 0 paper-tested · 8 evidence gaps listedOpen the radar →
Physical AI watchlist

Who supplies the joints of humanoid robots

Novanta, Timken, Moog and USA Rare Earth on one evidence ladder: product fit, customer wins, production orders, humanoid revenue and margins. Confirmed facts, management targets, company-wide figures and unknowns are labelled apart, with dated company sources. Research priority, not buy ratings.

Open the watchlist →
Fund holdings (13F)

What funds disclosed, and what changed

Real 13F filings of funds named in investing reels, verified on SEC EDGAR: holdings, new and exited positions, and why a quarterly filing up to 45 days late is research, not a signal to copy.

Situational Awareness LP: positions as of Jun 30, 2026 · Altimeter Capital Management, LP: positions as of Jun 30, 2026Open the filings →
Theme exposure

Where Boostio's paper books overlap

The economic themes inside the house paper books (Wave's crypto book, Stock Trade Twins and the general stock picks), each book on its own, with anything unclassified shown as unknown.

Wave: 10 positions, 31% AI-linked · Stock Trade Twins: 4 positions, 0% AI-linked · General stock picks: no open positionsOpen the map →
Stock Trade Twins

Rival stocks and who leads now

Rival US stocks across every industry, which one is ahead of its twin, and a shared paper strategy that trades a gap only when it has a measured habit of closing.

Open Stock Trade Twins →
Stock Besties

Who depends on whom in AI

A sourced map of who funds, buys from, supplies and hosts whom among OpenAI, Anthropic, NVIDIA, Microsoft, Amazon, Oracle, CoreWeave and others, with the filing or release behind every link, the companies an event would reach and the exposure your picks share.

Open Stock Besties →
How to read it

Three kinds of statement, never mixed

1 · Facts

Filed and dated

Numbers companies and funds filed with the SEC, producer prices from the U.S. Bureau of Labor Statistics and final daily closes, each with its period, filing date and a link to the source.

2 · Hypotheses

Labelled as Boostio's

Why a company might benefit is Boostio's reading of the facts, shown next to its uncertainties and the specific data that would invalidate it. It is not a forecast.

3 · Paper results

Forward, after costs

An idea counts as paper-tested only once a simulated order has filled in a house paper book at a real opening price with modeled costs. No backtest is shown as a result, and nothing is proven until the record is long enough.

Boundaries

What these tools do not do

  • No broker connection and no real orders: every position here is simulated.
  • No copying of funds: 13F filings are delayed and incomplete, so they appear as research leads only.
  • No personal portfolios: the exposure map covers Boostio's own paper books, never anyone's real holdings.
  • No silent changes: the rules are versioned and dated, and a source that fails is shown as failed.

Follow the research as it updates

Boostio is a research-first platform in private beta: crypto candidates, daily market stories, public paper trading, rival stocks and now AI-infrastructure research. These research pages stay free and open.

Explore the private beta