Research journeys worth following.
Each case study starts with one question and shows the session end to end — the query, what came back, what is worth verifying, and what could go wrong. These are research starting points, not recommendations.
Follow the AI infrastructure chain
Ask about the trend everyone is already watching, then look one layer down.
The question
Which businesses supply the infrastructure that AI computing depends on?
The query
ai infrastructure bottleneck
Why it appeared
The matches rank on how each company maps to the theme, what the business actually does, and how the market is pricing it — so the chips, power, cooling, and networking layers surface side by side.
What to verify
Whether the exposure is material to the business, how that business is performing, and whether the story is already in the price. Our market note of 7 July 2026 recorded exactly this problem: a "bloodbath in semiconductors" while NVDA held, and the AI trade "compressing from a broad tide to a narrow moat".
What could go wrong
A theme that stops moving, a company with only tangential exposure, or a valuation that already assumes the best case.
From a Relaxfolio research run. A research starting point — not a recommendation or a forecast.
A theme that has already turned
The same tools that surface a theme will show you one that is unwinding.
The question
Where have the crowded themes gone — and which names sit there?
The query
rare earth stocks
Why it appeared
Ranking the theme by market data makes the turn visible: the 7 July 2026 note logged "Rare Earths -33%" over a month, alongside solar, batteries, and drones, with "little sign of capitulation".
What to verify
Whether the drawdown has changed the underlying business, or only the price — and whether the theme has a catalyst rather than a headline.
What could go wrong
Catching a falling knife. A theme that is down sharply can keep falling; a screen describes the market, it does not call a bottom.
Illustrative of the workflow. The theme performance figure is from our own market commentary — see Market Commentary for the dated note.
A cheap stock is not automatically a bargain
A low multiple is a question, not an answer.
The question
Which companies look inexpensive on the numbers — and deserve a closer look?
The query
low debt buybacks
Why it appeared
The candidates rank on balance-sheet strength and shareholder returns — the factors the question named, applied across thousands of US names.
What to verify
Whether the business is still generating cash, whether the debt is genuinely manageable, and whether the cheapness reflects a temporary discount or a permanent problem. A multiple is the beginning of the investigation, never the end of it.
What could go wrong
The value trap. The cheapest company on the screen can be the most expensive mistake if the business is deteriorating, the debt is binding, or demand is falling away.
From a Relaxfolio research run, using public filings and market data. Not investment advice.
How to read these
Are these recommendations? ▾
Are the results guaranteed to be typical? ▾
What data do they use? ▾
Why publish them at all? ▾
Start with a question of your own.
Every journey here began with one plain-English question. Yours can start the same way.
Run a query