Research for owners of compounding businesses.

The question is asked as if it had a yes-or-no answer. It does not. What we see is a set of very different assets growing at rates public markets have not recorded before, priced by a market that is paying partly for that growth and partly out of fear of being disrupted. Some of those prices are exaggerated. Some are cheap. The average of the two tells you nothing about either.

AI is the largest capital deployment cycle since the internet build-out of the 1990s. High valuations are what that flow of capital looks like from the outside. Our job is to take the question apart one company at a time: how much of today’s price is growth you can trace to cash flows, how much is fear, and how much is neither. Separating signal from noise inside that volatility is the whole of our work.

We do not measure risk by volatility. A stock that fell 30% in a quarter did not become riskier that quarter. The thesis behind it either held or it did not, and only the second case is a loss. We measure risk by thesis, on three things:

We would rather own assets that appreciate and grow than hold claims on money, and we think most readers would too. A bond promises to pay you back in a currency whose supply nobody controls. To know what that promise is worth, you need an inflation rate, and the official rate is the average of a basket that is not yours. You have to estimate it yourself. We do, and we hold it against every yield we are offered.

Among assets that grow, we prefer technological growth and digital scale: businesses whose next unit costs little to produce, whose distribution is global from the first day, and whose returns on capital rise rather than fall as they get larger.

Money supply risk sits inside our theses and our valuations, not beside them. We follow it and we measure it. Most value investors treat it as background. We treat it as a variable, and it is one of the things that makes our work different.

The holding period we write for is forever. Our measuring stick is a three-year internal rate of return. The two do not conflict. The three-year thesis is what we underwrite and what we can be held to. The quality work, on moats, management and returns on new capital, is the insurance that lets us roll that three-year thesis forward, again and again, for as long as it keeps being true.

Warren Buffett got the fundamentals right. Buy businesses, not tickers. Price matters. Hold through the noise. Stay inside what you understand. We build on all of it.

What he saw late, and said so himself, was the scale of value that global platform, digital and technology companies would create. Over the last thirty to forty years they have outperformed almost every other asset class available to a public investor. Artificial intelligence extends that record: capital deployed into compute, models and distribution is earning returns at rates that were not imaginable when the value-investing playbook was written. Our work is to apply that playbook to those returns.

CoreValue Research is led by researchers. Our job is to produce intelligence for buy-side portfolio managers, not to run their money. Discretionary management sits elsewhere, at CoreValue Capital, and that is also what finances this research. The separation is deliberate: it is what allows us to publish a conclusion without a book to defend, and to write about every company on the same terms, whoever happens to hold it.

Independent does not mean distant. We are in the loop every day: the morning news before the open, every conference call of every company we cover, and every event and interview we can reach. We also go to the people actually deploying these technologies, through our own proprietary expert network, because the engineers and buyers installing this equipment see the demand long before it reaches a quarterly report. Nothing on this site is published without one of us having read the filing it rests on.

We use artificial intelligence intensively, and we would rather say so plainly than leave a reader wondering. Not one assistant but a set of agents, working the way very fast, very literal analysts would: they read filings line by line, rebuild our data from the source documents, chart what they find, and hold every draft against the evidence underneath it. They are the reason a house this size can cover what we cover, at the depth we cover it.

The utility we see is not speed for its own sake. The hard part of this work was never the reading. It was carrying what we see out of our own heads and onto a page someone else can act on, in the time the information is still worth something. That distance has collapsed, and we think that is transformational, for this house and for the profession.

Above all, the agents are instructed to argue against us. Every thesis is put to a model whose brief is to find the evidence that breaks it, before we publish rather than after.

There is a second reason we use them this deeply, and it is the one that matters most for what you read here. We cover the companies building this technology. Using it every day, at the edge of what it can do, is part of how we form a view on what it is worth: our opinions on AI rest on being heavy users of it, not on watching it from outside. Wherever technology can be added to the job of an analyst, we add it. The judgment about what a business is worth, and what would have to be true to pay that price, stays with us. And none of it publishes itself: nothing goes up without one of us having read the filing underneath it.

We believe people should earn their money from returns on capital, not only from their labour. The practical consequence is that most people would be better served by more exposure to equity, in whatever form suits them: direct ownership, funds, or a stake in the company they work for. Independent research is what makes that exposure a decision instead of a bet. That is why we write, and why we publish our numbers, our sources and our mistakes.

We write for two readers at once: the buy-side portfolio manager who needs the work underneath a position, and the individual who wants to own good businesses and be able to say why. The second reader is the reason we explain our terms. The first is the reason we never round them off.