Actionable insights — I'm Buying More Today

4:36 1. Trade the serial re-rating: buy the peer group's last holdout

1. Define a tight peer group by a shared controversy, not by sector — here, the four companies whose enormous capital spending the market distrusts for the same reason.
2. Establish that the market resolves that controversy serially, name by name, on earnings evidence — not simultaneously for the whole group. Look for one member that has already been re-rated as proof the pattern exists.
3. Rank the remaining members by how much of the doubt is still priced in. The one with the most unresolved skepticism is the position to add to; the ones already re-rated are held, not chased.
4. Wait for the resolving event to be an earnings report (a scheduled, evidence-bearing catalyst), and size the position so you can hold across several of them.

• A peer group sharing one bear objection; at least one member already re-rated on evidence; a remaining member whose objection is untested — plus an honest read of why it's last (Meta has no public cloud where the spend becomes billable revenue, so its proof takes longer and the trade needs more patience).

10:41 2. Decompose an EPS miss into one-time items before believing it

1. Treat a headline EPS miss as a question, not an answer: is this the earnings power declining, or accounting noise? "In some cases, earnings per share misses… are a sign of a declining or troubled business. But there's also cases where that's not what's happening."
2. Go to the cost-and-expenses line of the actual report and itemize what the analyst models did not contain — legal settlements, restructuring/severance, impairments, and changes to depreciation schedules.
3. Add the non-recurring items back and re-ask whether the underlying number was in line. Separately flag accounting choices (a faster depreciation schedule than peers) as conservatism, not deterioration.
4. Cross-check the operating lines that can't be dressed up — revenue growth, users, engagement, margins. If those are intact while only EPS missed, the miss is composition, not condition.
5. Apply the same skepticism in reverse: strip gains too (Amazon's EPS "a lot of that is equity stakes… we'll ignore for now") and judge the operating beat.

• A miss driven by items named explicitly in the cost lines rather than by weak revenue; a depreciation-schedule change disclosed alongside a capex ramp; operating metrics still accelerating while EPS falls — the signature of a decomposable miss.

3:53 3. Split the theme: own the demand aggregator, cap the scarcity supplier

1. For any boom, ask where each company sits: is it selling into the shortage, or is it the party creating the demand and controlling the end customer?
2. Test the seller's volume for durability: how much is one-time, scarcity-driven ordering that reverses when supply catches up? Commodity-like economics plus scarcity pricing means both price and volume can fall together.
3. Test the buyer for aggregation: does its spending buy an infrastructure layer it will own, with distribution to the end user — i.e. will it be the "demand aggregator" that captures the pricing?
4. Allocate accordingly: keep a leg in the best supplier for exposure, but concentrate capital in the aggregators. Refuse to concentrate in the supplier category regardless of how good the current numbers look.

• Order books that are pull-forward rather than run-rate; supplier margins that depend on shortage pricing; a buyer whose capex converts into an owned layer with end-customer distribution — the tell for which side of the theme keeps the economics.

11:54 4. Score management's prior explanation before extrapolating a scary metric

1. When a headline metric breaks trend, first read what management said caused it — and whether they gave a specific, falsifiable reason and a forecast for the next period.
2. Refuse to extrapolate a single data point into a secular story until the next print. A one-quarter reversal with a named external cause is a very different object from a trend.
3. Wait for the next report and grade the explanation. If management called it accurately, raise your weighting on their future disclosures — and lower it on the outlets that called the disclosure deceptive.
4. Keep the scoreboard: narratives that were wrong once about a company are usually wrong the same way again, and rarely get retracted.

• A specific, checkable cause given at the time of a bad print; the following quarter's data confirming or refuting it; media framing that upgrades a one-quarter dip into "the end of the category" — the setup where the narrative, not the business, is the mispricing.

18:49 5. Invert the bear case — assume it's true and see what it proves

1. First check the bear's factual claim against primary sources (the earnings report, the call transcript) rather than the interview. A claim that a company "never said X" is checkable in one search.
2. Then grant the claim for argument's sake and follow it to its conclusion: if the disputed spending really isn't driving the growth, what does that imply about the core business and about the spending?
3. If the bear case, taken at face value, implies a better setup (an unaided core compounding fast, and a giant discretionary cost that could be switched off), it isn't a bear case — it's an unpriced call option.
4. Separately, discount demands for a metric that can't exist: if a benefit is diffuse and product-embedded, no honest management can attribute an exact dollar figure to it, so its absence is not evidence of absence.

• A bear thesis resting on a claim the primary document contradicts; a demand for a disclosure that is structurally unmeasurable; a bear case whose true version would flatter the business — all three mean the objection is noise you can buy against.

19:57 6. No leverage is what licenses patience — and concentration plus leverage is the standing failure mode

1. Before underwriting a "the market will come around" thesis, ask what could force you out early. The answer is almost always borrowed money or a redemption clock, not the business.
2. Structure so time is on your side: unlevered positions, sized to be held through several disappointing prints, so a 22% drawdown is an adding opportunity rather than a margin problem.
3. Run the post-mortem in advance on the opposite structure: concentration into a handful of scarcity-priced bottleneck names plus borrowing. Leverage magnifies both directions, so meteoric first-half returns are evidence of fragility, not skill.
4. Treat messianic, "only a few of us can see it" writing as a live risk indicator on a manager — hubris precedes the risks that "they don't need to take."

• Borrowing used to magnify a concentrated position in the year's hottest scarcity trade; a manager whose public writing claims unique foresight; and, on your own book, any position whose thesis needs years but whose funding needs quarters.