Actionable insights — Weekly SSR: the AGI earthquake overreaction, the FPH land-NAV screen, ServiceNow's AI-enabler test, rate-peak income adds, the data-center-delay sellers' market

1. Size the disaster against the damage — the "market-cap destroyed vs dollars lost" overreaction screen

1. When an operational shock crushes a stock, don't react to the headline — quantify the actual economic damage and compare it to the market cap erased. Convert the operational hit into dollars: lost/deferred production × price, plus elevated costs and repair capex.
2. Localize the damage on the asset map: is the hit to a core asset or a peripheral one? Pull the NAV split and confirm whether the crown-jewel asset (the bulk of NAV) is affected at all.
3. Derive the warranted drawdown (damage ÷ enterprise value) and compare to the realized drawdown; the gap is the tradable overreaction. Cross-check sell-side resets — a price-target cut that still implies a near-double tells you the panic, not the fundamentals, moved the stock.
4. Buy in tranches across listings/sessions (e.g. start in the home-market shares, add the US line) to average into the dislocation.

• Disaster-driven single-session plunges; the NAV share of the impaired asset; analyst PT cuts that still imply large upside; automated/algo-driven cascade selling decoupled from the real damage.

2. Value the land bank on marked NAV, not earnings — see through the JV and the lumpy print

1. When a company is structurally mis-bucketed (a landowner priced as a homebuilder), reject the P/E entirely — earnings are intentionally lumpy because land-sale timing is discretionary, so trailing multiples mislead.
2. Find the hidden value: the best asset is often held in an unconsolidated JV, so it never shows in consolidated revenue — only in periodic distributions. Look for a history of large lump-sum payouts as proof of the cash-generative reality behind quiet quarters.
3. Build the real enterprise value correctly: count all share classes (A and B) to get the true equity value, then compare to the marked real-estate NAV (accounting inventory marked to current value), not book.
4. Express the gap as price-to-NAV and size it to the liquidity: a 3x to NAV is only worth a small allocation if the name trades a couple-hundred-thousand shares a day.

• Asset-rich names mis-classified by sector; unconsolidated JVs holding the crown jewel; a record of lumpy distributions; multiple share classes that distort the headline market cap.

3. Separate the AI victim from the AI enabler — the cheap-software test

1. Start with the dislocation screen: a business whose free cash flow is rising fast while its multiple is falling hard (here FCF +36% over 12 months vs a −60% valuation) is being sold on narrative, not numbers — flag it for the moat test.
2. Run the AI-exposure question explicitly: does AI replace this product, or does the product orchestrate and govern AI? A single-data-model platform that becomes the control tower for agents — built across all hyperscalers, model-agnostic, governing third-party agents — is an enabler, not a casualty.
3. Sanity-check the durability with third-party confirmation (a sell-side desk reaffirming the AI-ACV and revenue path) and the customer base breadth (thousands of enterprise accounts across industries/geographies).
4. Anchor the buy to a DCF/FCF-yield floor above the index and add on a catalyst (post-FOMC weakness), expecting the re-rate when the "AI contagion" fear fades.

• FCF-up / multiple-down divergences; whether the product governs vs competes with AI; sell-side reaffirmation of AI revenue targets; an FCF yield comfortably above the S&P.

4. Add the most rate-hated, highest-yield income at the cyclical rate peak

1. Form the rate view first: when you judge long rates near a cyclical peak (here on a peace-deal/inflation-receding thesis), the most rate-sensitive income names are the cheapest and pay you most to wait.
2. Go where the yield is highest and the rate-beta is greatest — leveraged mortgage REITs (14–16% dividends) — but treat them as a basket to diversify single-name book-value risk, and make them core-sized (vs a speculative torque name).
3. Scale in over time ("we'll continue to add") rather than all at once, because the rate-peak call is probabilistic and the open can stay soft.

• Your own conviction that rates are peaking; mortgage-REIT dividend yields at extremes; book-value trajectory; sizing the basket vs the speculative sleeve.

5. Buy the sentiment extreme into a structural bid — the contrarian commodity add

1. Quantify the wash-out: a ~25% drawdown from the highs plus a positioning gauge at a multi-year bearish extreme (here the highest bearish gold positioning since 2017) is the contrarian setup.
2. Confirm a structural buyer that doesn't care about the dip (central-bank buying "not going away") so the fundamental floor is independent of the speculative selling.
3. Express it in layers to spread entry risk: spot ETFs for clean exposure, miner ETFs for leverage to the metal, plus a single-name special situation for alpha on top — and accept you may be "a bit too early," so add after the catalyst (post-Fed), not only before.

• Drawdown depth + positioning extremes (COT/desk data); an inelastic structural buyer; layering spot + miners + a single-name; adding after the macro event rather than only ahead of it.

6. Hedge a single-name long through the unlock — paired short + covered calls

1. Map the unlock calendar of the correlated asset precisely (each tranche, each date, the share count vs current float) to know when supply hits and how large it is relative to the float.
2. Hedge the long with the most-correlated, most-liquid leg — short a fraction of the exposure in the newly-listed underlying — so a complex-wide selloff is partly absorbed by the short.
3. Layer in income: sell rich, longer-dated covered calls against the long to harvest the elevated premium the locked-up holders' hedging creates — the premium leak further cushions drawdowns.

• The full unlock schedule vs float; correlation of the hedge leg; covered-call premium richness from forced hedging; the August/November supply dates as risk events.

7. Read the buyback that doesn't work as a sector sentiment gauge

1. Find the extreme capital-return case — a company retiring a huge share of its float (even debt-funding an accelerated buyback) — and observe the price reaction. A buyback that should move a stock but doesn't is information.
2. Read the non-reaction as a sector verdict, not a company one: if even tens of billions of repurchases can't lift the stock, the whole sector is being sold for a structural fear (here AI terminal-value risk) and the flow is going elsewhere (semis — "one of the most crowded trades of all time").
3. Use it as a contrarian thermometer: peak disinterest in a cash-generative sector is the setup, but wait for the fear to be "fully understood" before pressing the long.

• Accelerated/debt-funded buybacks with no price response; sector-wide vs single-name explanations; the crowded opposite trade absorbing the flow.

8. Verify a shortage thesis on the physical bottleneck — and buy whoever already has the capacity

1. Triangulate competing research instead of taking one number: when two banks claim 70% of builds are stalled and a specialist pushes back to "under 50%," the actionable conclusion is the overlap — delays are large regardless of the exact figure.
2. Anchor the thesis on physics, not capital: track the hard constraints (transformer lead times 12–18 → 30–36 months, up to 5 yrs; grid-interconnect queues 3–5 yrs; capacity clearing prices +1,037% to $329/MW-day) — these can't be fixed with money.
3. Invert to the winner: if new supply slips, whoever already has live capacity and secured power sits in a sellers' market with premium pricing — own those operators while the bottleneck persists ("at least a year"), even if the longer-term picture is a bubble.

• Transformer/interconnect lead times and capacity clearing prices; ERCOT battery-cancellation data (a forward indicator); which operators have power secured today; the gap between announced and broken-ground capacity.

9. Track the pricing-power inflection when the dominant buyer swaps inputs

1. Watch the largest customer/backer of a "moat" provider: when even it diversifies away (adding a far cheaper open-source model and moving to usage-based, metered pricing), the incumbent's pricing leverage — and the "moat" — is structurally breaking.
2. Quantify the cost gap that forces the shift (here DeepSeek ~1/3 the input / ~1/7 the output cost of Claude) and corroborate with other large buyers balking (banks blocking the model on cost/compliance).
3. Trade the second-order exposure rather than the unlisted principals: short the listed wrappers/funds whose value depends on those private "moats" staying intact, and note the bull case (metered pricing → more total AI consumed) that limits how far you press it.

• The anchor customer diversifying suppliers; usage-based replacing flat-rate pricing; cost-per-token gaps; listed funds with concentrated private-AI secondaries as the shortable expression.

10. Trade the binary geopolitical catalyst — keep a cheap hedge while it resolves

1. "Rates drive the world and the war drives rates." Map the on/off catalyst to the macro chain in both directions: peace → oil/dollar/rates down → rate-sensitive value up; breakdown → oil up → inflation/rate fear back on.
2. Don't bet the binary outright — carry a small, cheap hedge on the tail you fear (long a little oil ahead of a possible peace failure) so a reversal pays rather than hurts.
3. Keep dry powder for the dislocation the headline creates: a renewed oil spike or risk-off open is the buy window for the names you already want (gold, the dividend basket), so pre-commit the shopping list.

• Oil/dollar/10-yr reaction to each headline turn; a cheap option/hedge on the feared tail; a pre-set buy list for the risk-off open; the PCE print as the next structural-inflation check.