Investment research through a long-term owner's lens

NVIDIA invests in OpenAI, OpenAI signs hundreds of billions of dollars in compute purchases from Microsoft, Oracle, and CoreWeave, and those buyers turn around and buy NVIDIA chips — bulls call this loop a virtuous circle, bears compare it to Cisco in 2000. After checking every deal against primary documents, our verdict: the cash genuinely moves and SEC disclosure reaches the contract level, so the Ponzi label fails; but NVIDIA's top three customers already account for 54% of a single quarter's revenue, and the credit risk of vendor financing is concentrating in the chain's anchor. We put the subjective odds of a contract-level crack among the loop's core parties before the end of 2027 at roughly 65% — and since June the market has already begun pricing the two ends of the loop apart.

First edition of our quarterly Growth Score Index, built from 897 fully scored companies in the Zen Horizon research library. The median score is 43/100 and exactly one company clears 65, because the test is a credible ten-year 5x path, and durable quality turns out to be far more common than durable hypergrowth. Twelve names combine a top-decile score with a price below our fair-value band.

In early June, Micron fell hard from its all-time high, dropping as much as 13.3% intraday in a single session and pushing the question "has the AI memory supercycle topped?" to center stage. This piece uses our 71 semiconductor Baillie Ten-Question scorecards to make a contrarian call: the demand supercycle is real, but after Micron and Hynix rose roughly tenfold and crossed a trillion in market cap, the stocks have been priced to perfection. None of the 71 scores above 3 on either the expectations-in-the-share-price question or the what-the-market-is-missing question. The verdict: this is valuation reverting to cyclical fragility, not demand collapsing, and the most dangerous trade is buying the dip. We put the subjective odds of Micron drawing down at least 25% from its June 3 high within 12 months at roughly 60%.

When a big company joins the S&P 500, index funds are genuinely forced by the rules to buy it. The "inclusion premium" that used to come with that, though, has all but vanished over the past two decades. Drawing on the membership rules of the three major index families, the mechanical buying from $20 trillion in passive money, and first-hand academic evidence on the "disappearing index effect," this piece asks what a spot in an index is really worth. The verdict: passive demand equals roughly 7–8% of a new member's float and is genuinely measurable, but the excess return from announcement to effective date has fallen from 7.4% in the 1990s to 0.3% over the last decade. SpaceX cannot get into the S&P 500 and is about 15 trading days from the Nasdaq-100; its rock-bottom 4.25% float is the real variable.