Additions 2026-09-23 — HHH (Howard Hughes Holdings), industry stages, 2026-09-23
Cluster pass, 23 Sep 2026.
Additions 2026-09-23 — HHH (Howard Hughes Holdings), industry stages, 2026-09-23
Slot cluster-additions-2026-09-23, member HHH only, split into its own file as the slot note allows ("Split into per-name cluster files if that is cleaner"), following _clusters/cluster-additions-NLY-2026-09-23.md. Run in a session on the owner's request, not by the routine. Model-drafted; nothing here is human-verified. Every figure carries the filing it came from; [background] marks background knowledge, [search-summary] marks anything that came from a search result, and transcript content is paraphrased and never put in quotation marks. 1913.HK is not scored here and is being run in parallel by another session.
Stages: S0 (lens), S4, S6, S7. Instrument: the standard AI-surplus four (holding_co, as the slot note directs), plus — because HHH now underwrites — the SKILL.md §4.3 risk-carrying variant as a second reading on the Vantage segment only (§4.1 re-tag rule; §2 below says why it is a second reading and not a re-tag).
§0 — Sources and how quotes were taken
Rung 1 (SEC EDGAR, User-Agent sent) answered everything filing-based: the FY2025 10-K (accession 0001628280-26-009710, filed 2026-02-19), the FY2025 10-K/A (Part III only), the Q2 2026 10-Q (0001628280-26-053324, filed 2026-08-05), the 2026 DEF 14A (0001104659-26-098808, filed 2026-08-19), the 8-Ks of 2025-05-06 (Pershing Square share purchase and services agreement), 2025-12-18 (Vantage purchase agreement), 2026-02-04 and 2026-02-17 (notes refinancing), 2026-04-21 (Grandisson warrants and board seat), 2026-06-05 (Vantage closing and the Series A preferred), the S-3ASR of 2026-08-14, and XBRL companyfacts for CIK 1981792 and the predecessor issuer, The Howard Hughes Corporation (CIK 1498828). Equibles (rung 4): Q1 and Q2 2026 call transcripts and FINRA short interest, 3 calls. 3spread (rung 4b): holders, owners, insiders, sales. ROIC.ai (rung 5): prices, 1 call. Two WebSearch queries (rung 6).
Each filing was converted to text (tags stripped, entities unescaped, whitespace collapsed), and every quoted span below was located in that text by exact substring match before it was written, then re-checked by script before the file was saved. Transcripts are Equibles machine speech-to-text with visible errors ("Mark Randison", "Ferdinand Square", "Vintage") and are paraphrased only.
§1 — The finding: HHH now owns about half of an insurer whose float Pershing Square is moving into stocks
Everything the market knew about HHH until 2025 was a land business: master planned communities (MPCs) in Las Vegas, Houston and Phoenix, condominiums at Ward Village in Honolulu, and the commercial property those communities need. The 10-K describes the land bank as "one of the nation's largest portfolios of MPCs, spanning approximately 101,000 gross acres across five states" (FY2025 10-K, Item 1). That business still produces all of HHH's earnings history.
Since 2026-06-04 it is also something else, and the filings say so in three places that have to be read together:
- HHH bought a specialty insurer and reinsurer. Vantage Group Holdings, Bermuda, for $2.1bn cash; at 2026-06-30 its assets are $6.30bn of HHH's $15.91bn (40%), with $2.12bn of claim reserves and $1.41bn of unearned premium now on HHH's balance sheet (Q2 2026 10-Q, segment asset reconciliation and balance sheet). Vantage's own annualised scale is roughly $1.2bn of trailing net written premium on $1.8bn of book (Q2 call, Grandisson, paraphrased) — against HHH's 2025 total revenue of $1.47bn (XBRL).
- HHH common does not own all of it. The purchase was financed with cash and "$1 billion of non-voting preferred equity financing from Pershing Square Holdings, Ltd. (“PSH”)." (8-K 2026-06-05). That preferred exchanges into units of the insurance holding company at a fraction equal to $1.0bn over the purchase price plus any capital HHH later adds — 47.6% on the purchase price alone ($1.0bn / $2.1bn, computed from the 8-K formula), capped at "more than 49% of the total shares of Buyer Units outstanding" (8-K 2026-06-05). HHH's call on it is priced at the greater of 4% accretion and "1.5 times the book value of Buyer (excluding non-controlling interests and good will and purchase-related intangibles attributable to the completion of the Transaction)" (same). In words: if Vantage does well, the call costs 1.5x its tangible book; if it does badly, the preferred shares the loss pari passu. HHH common's economic share of Vantage is therefore a little over half, and the other half belongs to a Pershing Square fund.
- The float is being moved into equities managed by Pershing Square. On closing, Vantage's subsidiaries signed investment management agreements "pursuant to which Pershing Square will act as investment manager of Vantage’s general account and other investment portfolios", and "As long as the Services Agreement remains in effect, none of the acquired subsidiaries will pay any additional investment management or advisory fees under the separate investment management agreements." (Q2 2026 10-Q, Note 3). Within 26 days the $2.6bn acquired bond book was sold ($2.34bn of proceeds, Note 4) and $1.08bn of common stocks bought — 32% of Vantage's $3.39bn portfolio at 2026-06-30, with "approximately 78% of the aggregate fair value of our investments in equity securities was concentrated in seven companies." (10-Q, Item 3). By the Q2 call the CIO put equities at about 40%, with a stated destination of somewhere north of 50%, and the net reserves backed by short Treasuries plus a cushion (paraphrased). The company's own sensitivity: a 10% equity move is ±$107.8m pre-tax (10-Q, Item 3) — against HHH's H1 2026 net income to common of $166.6m.
Put together: HHH common owns about half of a five-year-old insurer that S&P cut to BBB on the change of ownership (§3), whose asset side is now a concentrated Pershing Square equity book run without an additional fee, and whose other half is owned by a Pershing Square fund. Pershing Square is on four sides of the structure — 46.6% common holder (2026 proxy), fee-earning adviser to HHH (8-K 2025-05-06), preferred holder in the insurer (8-K 2026-06-05), and manager of the insurer's float (10-Q Note 3). The committee of independent directors approved the financing under the standstill (8-K 2026-06-05). None of that is hidden; it is the design. The question for S7 is whether a shareholder outside Pershing is paid for sitting in it.
§2 — HHH · holding_co · ✅ lens applied — standard 3/4; §4.3 second reading on Vantage 0/2 applicable (2 n/a)
Which instrument, and why not a re-tag
§4.1 says re-tag a mis-clustered member before scoring. HHH is not mis-clustered: it is a holding company that now owns a land business and roughly half an insurer. The land business is still the larger part on every measure the filings allow — the company's own conservative value estimate put about 80% of today's value in real estate and 20% in its economic share of Vantage (Q1 call, CIO, paraphrased; a company claim, not reconciled here); all the earnings history is real estate; and HHH's net investment in Vantage after the $1.0bn preferred is roughly $1.1bn plus the capital added since (the Q2 call mentions a $300m contribution, paraphrased) against $3.96bn of common equity. The executive chairman said on the Q2 call that the intent is to become disproportionately an insurance holding company over time (paraphrased). Recommendation to the parent: keep holding_co; do not re-tag to insurance now. Add a note to the watchlist row that S0 carries a §4.3 second reading on Vantage, and re-examine the tag when the FY2026 10-K reports Vantage as a segment with a loss-development table, or when Vantage exceeds half of common equity.
Does the MPC business fit the standard test? Better than a pure developer would, because the test's categories include "a permitted quarry or landfill" and entitled land in a supply-limited metro is that shape. Test 2 is the one that does not fit, and why is itself a finding (below).
Standard AI-surplus four (account skill wording, not reworded)
| # | Test | Score | Evidence |
|---|---|---|---|
| 1 | Distribution lock-in | 1 | Structural, from ownership: inside an MPC HHH is the only seller of residential land, and a homebuilder that wants lots in Summerlin, Bridgeland or The Woodlands buys them from HHH — 412.3 superpad acres in Summerlin alone in 2025 (FY2025 10-K MD&A). The company describes managing lot supply to builders against underlying home sales in each community (Q1 call, CEO, paraphrased). For Vantage: brokers are the channel, "Willis Towers Watson 11.9 % Howden 10.3 %" of post-acquisition gross premiums (10-Q, Note 1); the account skill names broker distribution as the specialty-insurance lock-in, but a five-year-old carrier's broker relationships are not yet a book the broker cannot move — the MPC carries this score |
| 2 | Unscrapeable data | 0 | Not the absence of the word "data" — the series is thin by the company's own account. For the insurer: "Its loss-development patterns, pricing assumptions, reserve estimates, and underwriting models are based on a thin historical data set as compared to longer-tenured insurance carriers" (10-Q Item 1A; Vantage "was founded in late 2020"). For land: the price series that matters (lot and home sales) is transacted in public, disclosed by HHH itself each quarter and recorded by the county [background]; the daily tracking of home sales per community is useful operations, not a data asset a model cannot reproduce. What differs from a Tier 1 reference insurer: those have decades of loss experience, this one has five years |
| 3 | Liability absorption | 1 | The insurer's product is risk onto the balance sheet: $2.12bn of claim reserves, $0.56bn recoverable from reinsurers (10-Q). The land business absorbs entitlement and horizontal-infrastructure risk and is repaid through municipal utility district (MUD) and special improvement district (SID) receivables — $579.2m of MUD receivables at 2026-06-30 (10-Q balance sheet); the builder buys a finished superpad |
| 4 | A bottleneck tokens can't manufacture | 1 | Entitled, serviced land in supply-limited metros: ~101,000 gross acres (10-K), including 37,000-acre Teravalis west of Phoenix (2026 proxy, CEO achievements). Ward Village entitlements were widened by the State of Hawai‘i by an estimated 2.5–3.5m gross sq ft (10-K Item 1). For the insurer, licences and ratings: goodwill is attributed partly to "financial strength ratings that could not be replicated on a comparable timeline through internal development" (10-Q Note 2) |
Standard lens 3/4 (all four applicable). Structure, not cycle: the land price fell in 2025 — "412.3 acres sold at an average price of $970,000 per acre in 2025, compared to 216.5 acres sold at an average price of $1.3 million per acre in 2024" (10-K MD&A, Summerlin) — which belongs in the falsifier, not in test 4.
Surplus question — who keeps the surplus cheap intelligence creates? For land, the owner of the scarce input: if AI raises output and incomes in the Sun Belt metros HHH sits in, that shows up as demand for housing and commercial space in a place with a fixed supply of entitled land, and the land owner captures it as price (the classic rent argument [background]). The executive chairman also floated data-centre or company-town uses for the Teravalis land (Q1 call, paraphrased) — an option, not a plan. For the insurer, AI lowers underwriting and claims cost, which competition passes to insureds unless the carrier has an edge; Vantage's own chair expects AI to help execution and does not expect it to end the cycle (Q2 call, Grandisson, paraphrased). The land keeps the surplus; the insurer mostly passes it on.
Falsifier — one number. MPC EBT on a trailing four-quarter basis below $349.1m — the 2024 figure, before the "record" $476.1m of 2025 (FY2025 10-K MD&A). The whole land case is pricing power on a finite bank; a return to 2024's level without a housing recession would say the 2025 result was volume, not price.
Liquidity. ADV $41.6m/day (21 sessions 2026-08-24 → 2026-09-22, close × volume, ROIC NYSE:HHH). Last price $66.09 (2026-09-22). 59.72m shares outstanding (2026 proxy record date, 2026-08-17) → market cap ~$3.95bn; Pershing Square holds 27.85m (46.6%), so the float outside it is ~31.9m shares, ~$2.1bn. P/B 1.00 on 2026-06-30 common equity of $3.96bn ($66.39/share); 1.28x tangible book ($51.83/share after $282.2m goodwill and $586.4m intangibles, 10-Q balance sheet).
Second reading — §4.3 risk-carrying variant, Vantage only
Applies to: the insurance segment (not yet a reportable segment; the 10-Q reports it as "Vantage"). Size: $6.30bn of $15.91bn assets; about half of it economically HHH's (§1).
| # | Test | Score | Evidence |
|---|---|---|---|
| 1 | Reserve honesty | n/a | Not scorable from HHH's filings. The reserves were marked to fair value at acquisition ($2.09bn, 10-Q Note 2) and no loss-development table has been published; Vantage has five accident years and was private. The one quarter disclosed had $19m of adverse prior-year development, largely in a discontinued transaction-liability line (Q2 call, Grandisson, paraphrased; Vantage-standalone basis). The FY2026 10-K should carry the first table. n/a, not 0 — but the one data point is adverse |
| 2 | An edge that survives the cycle | n/a | Five years of ex-cat ex-development combined ratio are not in any filing read. The company's figures: Q2 2026 combined ratio 101.6% (94% a year earlier) with $18m of catastrophe losses and the $19m of adverse development; current accident year ex-cat 91.4% in Q2 and 90.9% for H1 against 94.6% (Q2 call, paraphrased, Vantage-standalone GAAP). The 26-day consolidated stub shows 95.2%, which the 10-Q itself says is not representative |
| 3 | Capital that is theirs, and rated | 0 | Premium-to-surplus is conservative at about 0.7 ($1.2bn trailing NWP on $1.8bn book, Q2 call, paraphrased) and A.M. Best is "A- (Excellent) with a positive outlook" (10-Q Item 1A). But S&P downgraded Vantage's operating unit to BBB from A- on the change of ownership, citing the new parent's credit, and then withdrew the rating at the company's request [search-summary] (The Insurer, Royal Gazette, Intelligent Insurer, 2026-06-15/16). The test's no-downgrade-in-three-years condition fails. Separately, AdVantage — a Bermuda segregated-accounts vehicle "utilized to access third-party investor capital and provide underwriting capacity for certain reinsurance business originated by the Company" (10-Q Note 1) — books fee income on risk carried by outside capital, and the stub ceded 31% of gross written premium ($156.1m gross, $107.2m net, 10-Q MD&A). Trend not observable yet |
| 4 | Growth that does not dilute | 0 | The insurer was bought with $900m of new HHH common sold at $100 (8-K 2025-05-06; shares 50.4m → 59.4m, +18%) and $1.0bn of preferred that takes up to 49% of the insurer's units (§1). The growth of HHH's insurance book is, so far, entirely rented from Pershing Square |
§4.3 reading: 0/2 applicable, 2 n/a — fewer than three applicable, so it cannot pass by construction. That is the finding: the insurer is too young and too new to HHH's filings to be scored on the two tests that decide an underwriter, and the two that can be scored both fail on structure (a rating cut by the change of parent; growth financed by outside capital).
Cycle question (§4.3). Measure: Vantage's current-accident-year ex-cat combined ratio, 90.9% for H1 2026 against 94.6% a year earlier (paraphrased). Grandisson places the P&C market mostly in the third stage of his four-stage cycle — rates off the peak but still adequate — with casualty stalled in stage 2 and some property lines entering stage 4 (Q2 call, paraphrased). The case does not need a hard market as much as it needs the equity market: Vantage is small enough to pick its spots, but at 40–50% of invested assets in stocks its book value moves with Pershing's portfolio.
AI line (one sentence, not scored). Cheap intelligence lowers Vantage's underwriting and claims-handling cost but is priced away by competitors in a soft market; it does not change who carries the reserves.
Falsifier for the insurer. Any adverse prior-year development in the FY2026 10-K table for accident years 2021–2025, or an A.M. Best action other than an affirmation.
§3 — S4 Industry & supply map
Land chain. Raw land (bought at scale decades ago; Teravalis bought about three years ago per the CEO, Q1 call, paraphrased) → entitlement and horizontal development (HHH; roads, utilities, drainage, financed partly through MUD/SID districts that later reimburse it) → superpads and custom lots sold to national and local homebuilders, with builder price participation above a breakpoint (10-K MD&A) → homebuyers. Beside it: commercial land and operating assets (office, multifamily, retail inside the MPCs; Operating Assets NOI $262.0m in 2025, 10-K) and condominiums at Ward Village (two towers 96% pre-sold, $1.5bn contracted; three in pre-development 66% pre-sold, $2.0bn, FY2025 10-K Item 1).
Where the pricing power is. With HHH inside its communities: it is the only seller of land there and meters supply to builders' home sales; the builders compete with each other for lots [background: the national builders HHH sells to are not named in the filings read]. The limit is metro-level competition from other landholders, which the 10-K lists as a risk factor ("we compete with other landholders"). The land price is ultimately set by the home price the end buyer can finance — which is why the first risk factor is a housing downturn.
Insurance chain. Insured → wholesale and retail brokers (Willis Towers Watson 11.9%, Howden 10.3% of post-acquisition GWP, 10-Q) → Vantage (U.S. admitted and E&S, Bermuda specialty and reinsurance) → reinsurers and AdVantage segregated accounts (third-party capital) → float, invested by Pershing Square. Pricing power sits with the broker and the market in a softening cycle; Vantage's lever is selection, not price.
Concentration. Geographic, by the company's own list: Arizona, Texas, Hawai‘i, Nevada and Maryland; Summerlin tied to Las Vegas tourism and gaming; The Woodlands and Bridgeland to the energy sector (10-K Item 1A). Two brokers above 10% at Vantage. No customer concentration disclosure beyond that.
Who wins share, from whom. Sun Belt MPCs have taken residents from high-tax coastal metros — the executive chairman's own framing on both calls (paraphrased); the filings record the demand as new-home sales and land sales, not as a migration claim. In insurance, share is moving from under-capitalised to well-rated carriers in a softening market; a BBB-at-S&P, A- at A.M. Best carrier is on the wrong side of that where S&P is the rating a buyer requires.
§4 — S6 Scuttlebutt
Board claim rows. None: grep of ledger/pipeline/gold/ for Howard Hughes and HHH finds no claim row (the one hit, in a Market Huddle transcript, is the string "Shhh"). A failed search is not evidence; the board simply does not track a show that has discussed it.
Outside evidence 1 — the S&P action on Vantage, [search-summary]. Trade press (The Insurer, 2026-06-15; Royal Gazette, 2026-06-16; Intelligent Insurer) report that S&P downgraded Vantage's operating unit to BBB from A- after the takeover, citing the acquirer's credit quality, then withdrew the rating at the company's request, while A.M. Best affirmed A- with a positive outlook. S&P separately upgraded HHH itself (to BB+, per an Investing.com headline, stable outlook tied to Pershing Square's long-term ownership) [search-summary]. On the Q2 call Vantage's chair described S&P's action as reflecting its group methodology rather than Vantage's standalone quality (paraphrased); the 10-Q discloses the A.M. Best rating and does not mention S&P in the text read. Nobody on the call asked about it.
Outside evidence 2 — the short side. FINRA short interest 3.54m shares at 2026-08-31, up from 2.64m at 2026-06-30 and rising five settlements running; 5.9% of shares outstanding but 11.1% of the ~31.9m shares Pershing Square does not own; 8.3 days to cover (Equibles). That is a meaningful short on the float that trades.
Outside evidence 3 — commentary, [search-summary]. The mini-Berkshire framing is everywhere (Seeking Alpha, Substack, AOL/Business Insider headlines); the skeptic's line in the summaries is the fee load to Pershing Square and the Berkshire-textile comparison. Not read beyond the summaries.
Price action, from the filings. Insider withholding transactions were priced at $89.53 (2025-11-30), $79.77 (2025-12-31), $80.04 (2026-02-05) and $63.05 (2026-04-01) (Form 4s); ROIC closes $67.93 (2026-08-20) and $66.09 (2026-09-22). Pershing Square paid $100 in May 2025. The stock fell about 29% between late November 2025 and April 2026, across the Vantage announcement (2025-12-18); the cause is not established here and is not inferred.
Product test / expert call: ⏳ OPEN (user). For HHH the useful calls are two: a Las Vegas or Houston homebuilder land buyer on lot pricing, and a Bermuda specialty underwriter on how Vantage's book is regarded after the S&P action.
§5 — S7 Thesis and framework test
Framework used: a plain holding-company read — sum of the parts, who controls the capital, and what the capital is paid (the AI/semis lens table from _clusters/ai-capex-2026-09-20.md does not apply). Bear case first.
Bear case. HHH has become a vehicle whose best asset — a finite, well-located land bank that generates cash — is being liquidated on purpose, with the proceeds redirected into an insurer that is five years old, half-owned by a Pershing Square fund, rated BBB by the only agency that re-rated it on the change of parent, and investing its float in a concentrated Pershing Square stock book. The executive chairman said the priority for every incremental dollar of free cash flow is Vantage (Q2 call, paraphrased); the company has not said what return that dollar has to clear, when asked twice (dodge log). Pershing Square is paid a base fee of $15m a year plus 0.375% a quarter of the price above $66.15 on 59.4m shares (8-K 2025-05-06), and exiting that agreement on a change of control triggers a make-whole of the remaining ten-year term. The public float holder takes the land risk, the underwriting risk and the equity-market risk, and the upside of Vantage above 1.5x tangible book is shared with the preferred. Short interest on the float is 11% and rising.
Bull case. The land bank is real and pricing-powered — MPC EBT $349.1m → $476.1m (2024 → 2025, 10-K), Summerlin land appreciating at roughly 15% a year over five years (Q1 call, CEO, paraphrased), condo pipeline over $4bn with about 78% under contract (Q2 call, paraphrased) — and HHH trades at 1.00x book with that land carried at historical cost. Vantage came with an Arch-built management team: Marc Grandisson (former Arch CEO) as executive chair, who bought $10m of HHH warrants struck at $100 with his own money (8-K 2026-04-21), and a CEO-designate from Arch (Q2 call, paraphrased). The insurer's current-year ex-cat combined ratio is 90.9%. Management states it will not issue equity at anything near current prices (Q2 call, paraphrased), and the company's own conservative value estimate is $104 a share (Q1 call, paraphrased; not reconciled here).
Variant perception. The market prices HHH as a real-estate company at book with a Pershing discount. The variant is that the unit of value is no longer the land: it is what Pershing Square earns on Vantage's float, because (a) the land is being converted into insurance capital on purpose, (b) half the float's equity return goes to the preferred via the exchange formula, and (c) the float's equity sleeve is already ~$1.1–1.4bn, heading for more than half of Vantage's invested assets. HHH common is, at the margin, a levered and fee-bearing claim on Pershing Square's stock picking, with a land bank as the collateral. If Pershing's large-cap book compounds at mid-teens, the structure works and the market has the discount wrong; if it does not, the land is being sold to fund it.
Kill criteria (specific, testable):
- Adverse prior-year development for Vantage in the FY2026 10-K loss-development disclosure (any accident year 2021–2025 developing adversely in aggregate).
- A.M. Best places Vantage under review or changes the outlook to stable/negative — the only remaining A- after the S&P action.
- Vantage cash plus short Treasuries below net claim reserves at any quarter-end (net reserves $1.55bn at 2026-06-30: $2.12bn less $0.56bn recoverable) — the company's own stated rule for the barbell; equities funding reserves would mean the rule gave way.
- MPC EBT on a trailing four-quarter basis below $349.1m (the 2024 level; the S0 falsifier).
- A primary equity issuance off the 2026-08-14 S-3ASR below the latest reported book value per share ($66.39 at 2026-06-30).
Time horizon. 2–4 years; the first hard evidence arrives with the FY2026 10-K (first Vantage loss-development table, first full year of the equity sleeve) in February–March 2027.
Conviction: ⏳ OPEN (user).
§6 — What this pass could not do
- No loss-development table for Vantage exists in HHH's filings yet; tests 1–2 of §4.3 are n/a for that reason, not scored 0.
- The Vantage supplemental and the Q1 valuation deck were not read (investor-relations documents, not SEC filings; gate-minimum). The $104 / $211 value figures and the $5.6bn land residual value are company claims heard on calls, not reconciled.
- The S&P action is known only from trade-press search summaries; the 10-Q text read names only A.M. Best.
- The equity sleeve's holdings are not named in the 10-Q (sectors only: communications $514m, consumer products $243m, financial $208m, technology $113m). The executive chairman pointed to Pershing Square's own portfolio disclosures (Q2 call, paraphrased); not read.
- The proxy states Operating Assets NOI of $276m in one place and scores the bonus on $261,985,000 (2026 DEF 14A, CEO achievements vs the incentive table); the 10-K says $262.0m. The difference is not explained in the filings read.