When a business changes faster than the narrative written about it, a gap opens between what the company is becoming and what the market still believes it to be. That gap is the investment philosophy of this firm.
Markets don't misprice businesses. They misprice narratives about businesses. That distinction is the foundation of everything we do.
Every multiple is the residue of a story the market has agreed upon — what a company earns, how fast it grows, who it competes with, how large it can become. These stories are not wrong. They are simply slow. When a business changes underneath its narrative — when the underlying economics shift in a direction the market has not yet absorbed — the price does not move with the change. Prices move when stories move. Stories move when enough capital has been exposed to a new interpretation, often enough, for long enough, to believe it. That process takes quarters. Sometimes years.
This is not a market failure. It is a structural feature of how information travels and how consensus forms. A business mid-transition looks broken on the metrics designed to measure the business it used to be. The old numbers flag deterioration precisely as the new economics are taking shape. The mispricing is durable because repricing requires a new narrative to harden — and the people whose job it is to write that narrative are the last to publish it.
The gap between what a business is becoming and what the market currently believes it to be is the only edge we know of that is not arbitraged away by the act of exploiting it — because businesses are always changing faster than the stories told about them. That gap is what we are looking for. Finding it, and holding it with conviction until the narrative catches up, is the investment philosophy of this firm.
The gap closes when a signal — a contract, a margin print, a clinical outcome, a regulatory decision — becomes impossible for the old story to contain. Our job is to be positioned before that signal, with conviction earned from work the market has not done. That requires time. We underwrite over three to five years and size positions to be held through the period when the market has not yet agreed with us.
We operate globally, outside the S&P 500. Not because we are indifferent to large-cap equities — but because a business followed by forty sell-side analysts has a narrative that updates continuously. Forty professionals reprocessing and reinterpreting every quarter. The gap closes fast. A business covered by three analysts in Seoul, Helsinki, or Warsaw can sustain a meaningful mispricing for years. That is where we work.
The belief also shapes how we find candidates. We do not begin with a screen — screens find what looks attractive relative to today. We begin with a world view: a forward picture of how geopolitical realignment, technological change, and economic transition are reshaping where value will concentrate over the next three to five years. From that view we derive themes. From those themes we hunt the businesses most likely to own significant value within them — before the market has priced that ownership in.
This approach accepts a particular discomfort. We will be early, and early is indistinguishable from wrong for a long time. We accept that cost because the asymmetry it creates — buying before the narrative has hardened — is the only source of return that justifies concentration. We run a concentrated portfolio because diversification is a hedge against ignorance. If you understand a thesis deeply enough to hold through discomfort, diversifying it away is the mistake.
The question every investment must answer is always the same: is this business changing in a way that the market has not yet priced? We interrogate that question three ways simultaneously — not sequentially, as a checklist, but as a system of independent verification. All three lenses must have a view before we have a thesis.
The interesting parts of a business are rarely in investor relations or the 10-K. They are in the documents that exist because something happened: regulatory complaints, supplier disputes, foreign-jurisdiction filings that pre-date US disclosures, the one sentence a regional manager said too plainly in the wrong forum. This lens rewards the person who keeps reading after the obvious source is exhausted — and who follows the thread even when it does not flatter the position they want to hold.
We were trained inside McKinsey's private-equity practice, where the standard is not building the model — it is understanding the business well enough that the model becomes obvious. This lens asks why a margin is moving, not by how much. Whether the capital allocation decision makes sense given what management says they are building. Whether the unit economics support the thesis or quietly contradict it. A spreadsheet that looks elegant has often replaced an understanding that was harder to come by. We build the understanding first.
Geopolitics realigns supply chains. Energy policy reprices industries. A new technology destroys a moat that took a decade to build. This lens asks whether the business is in the path of a force larger than itself — and whether that force is durable enough to sustain the thesis over the time we need it to. It is the lens that demands the most continuous learning. Every quarter brings a new regime, a new technology, a new set of constraints to understand from first principles. The work is never finished.
When the three readings agree, we have a thesis. When they disagree, we have something more useful: the specific question that needs to be answered before we can have one.
A firm that bolts AI onto a legacy research process gets a faster version of the same process. A firm that builds from AI as a working substrate gets something categorically different: the ability to operate at a scale of signal, a breadth of instrument, and a speed of synthesis that a small, concentrated team could not sustain without it. We did not build these tools because it was on trend. We built them because the belief in Chapter I made them necessary — and because each of the three disciplines in Chapter III makes a demand that human bandwidth alone cannot fully meet.
The result is a research infrastructure that gives a concentrated team the analytical reach of a fund many times its size — without the structural inertia and career-risk incentives that come with that size. We stay small deliberately. We build to see further deliberately. These are not separate choices. They are the same choice, made twice.
Below are the six instruments currently operating across the fund's research and portfolio management stack. Each has a specific job. None is a substitute for judgment. All of them extend the surface on which judgment can operate.
Every piece of information entering a thesis is classified by source type: HUMINT (primary conversations with management, suppliers, customers), OSINT (public documents, filings, press), SIGINT (market signals, order flow, positioning), FININT (Form 4, 13D/13F, short interest), and GEOINT (satellite imagery, AIS tracking, geospatial data). A position requires at least three independent intelligence categories in corroboration before it earns conviction sizing. That is not a heuristic. It is a structural constraint on how we build confidence.
A specialist research engine that applies a nine-step analytical framework to every drug program we evaluate: clinical science, probability of success, rNPV valuation, commercial diligence, and competitive intelligence. Produces IC-grade investment memos for every biotech and pharma position. Operates at the standard of a Baker Bros or RA Capital analyst, purpose-built for the fund's concentrated portfolio mandate.
A Monte Carlo simulation engine that stress-tests every position across four scenarios — base, bull, bear, and catastrophic — attributing expected return to specific business drivers rather than aggregate price targets. Translates qualitative thesis conviction into quantitatively grounded position sizing. Runs continuously as new information arrives, not only at thesis initiation.
Named for Vega and Theta — the two forces it hunts. A five-factor composite screener for options premium opportunities across the eligible universe: annualized yield, theta/gamma efficiency, probability-adjusted yield, volatility risk premium skew, and the Goyal-Saretto IV/HV signal. Used to generate yield on positions, acquire equity below market, and hedge near-term event risk. Runs on a live options chain feed.
Automated technical analysis covering RSI, MACD, Fibonacci levels, VCP (Volatility Contraction Pattern), and pivot recognition. Used to time entries and manage exit points within the fundamental thesis — never as a primary signal generator. Gradient informs the when. The other instruments determine the what and the why.
Grounded in Marko Papic's constraints-based framework: political outcomes are constraint-space predictions, not narrative bets. Given what actors cannot do — structurally, economically, politically — the range of what they will do narrows considerably. Atlas applies this methodology to every country or geopolitical exposure in the portfolio, producing investment-grade variant perception on political outcomes rather than headline-driven risk flags.
None of these instruments are permanent. We update them when we learn something new. The firm we want to build is one that keeps sharpening its capacity to see — not by hiring more analysts, but by building better instruments for the analysts we have. AI is not the strategy. It is the testament to it.
Even when the three lenses agree — when the journalist, the operator, and the trend scout all point in the same direction — four conditions send us back to the beginning. These are not judgment calls we weigh against the strength of the thesis. They are fixed stops.
We are looking for the gap between reality and story. If the story has already updated — if the re-rating is complete or well underway — the gap is closed. Late capital bears the same risk for less of the asymmetry. When the sell-side is writing about the inflection, we have already missed it.
The operator's lens requires conversation. A business we cannot speak with directly about why a margin is moving, or where the real constraint on growth sits, is a business we cannot underwrite to our standard. No access means no position, regardless of what the model says.
Preferred equity that converts before common. Dilutive instruments that grow with the share price. Structures designed to transfer the asymmetry away from the equity we hold. We have walked away from theses we believed in because the capital stack captured what we were underwriting for.
Revenue recognition that obscures the signal we are trying to read. Auditors who change without explanation. Restatements that rewrite history without apology. The philosophy is worth more than any single investment, and the philosophy requires trusting what we can actually verify.
Two people spent most of a decade noticing the same thing independently: that the funds they encountered were operating from assumptions about markets, time horizons, and portfolio construction that they did not share. Not wrong, necessarily. Built on different premises.
The belief in this document is not a marketing premise. It is the actual reason every decision this fund makes takes the shape it does — why we are global rather than domestic, why we concentrate rather than diversify, why we begin from a world view rather than a screen, why we hold through discomfort rather than trade out of it. The belief generates the mandate. The mandate generates the discipline. The discipline, applied consistently, generates the edge.
The deeper ambition is simpler than any theme: to build a firm that compounds not just capital, but the quality of its own thinking.
That requires staying genuinely curious — about businesses, about the world, and about whether the framework we use today will still be the right one in five years. It is why the firm is built the way it is built.