Quarterly market commentaries from Horizon Kinetics LLC (Murray Stahl / Steven Bregman) — a firm source, like Barron's or Goldman; James Davolos' individual appearances live under james-davolos. Long-duration compounding, hard-asset/royalty, exchange and bitcoin theses.
An index product they actually use — 7%+ yielding gas-pipeline MLPs with regulated inflation pass-through and no K-1, solving the wasting-asset problem bonds have under inflation.
ETF-Divide exhibit — 13.7%/yr per-share earnings growth since 2000 at 11x, having retired 58% of its shares in five years and shrunk itself out of index eligibility.
Same thesis as 2015 and "right on schedule" — cost of production doubles at the April 2028 halving to ~$130k implying ~$225k, cross-checked by a sub-t^6 Metcalfe power law at $270,438; expect ~33%/yr, not 90%.
Options exchange ahead of the S&P over 12 months despite the May selloff and cheapest in three to five years; pushing near-24/5 U.S. equity trading — expansion, not disruption.
Derivatives toll booth at a 17-year valuation low while it builds the 24/7 world itself (crypto, gold, oil futures, single-stock, prediction partnerships) and litigates to keep perpetuals inside the regulated perimeter.
Exchange/clearing toll booth at its cheapest valuation since 2009 — 15% 20-year per-share earnings growth, 33% after-tax margin, 17% of Polymarket for the data, and no futures contract failure in three decades.
The firm's own active ETF as a "completion fund" — its S&P-overlapping holdings are a 0.57% index weight, so the index carries almost no offset to commodity-price inflation.
Horizon-anchored July 2024 IPO — 300k+ Delaware Basin surface acres earning royalty-like water fees, with 4-to-6 barrels-per-barrel water cut and CPI escalators giving 12%+ revenue growth at zero capex, plus "powered land" data-center optionality.
Newest listed U.S. exchange (2025) and a disclosed significant holding — 40% revenue growth and 2x+ free cash flow in the March 2026 quarter, pre-scale economies; the "stunning decline" narrative is factually wrong.
Same anomaly with a smaller float — 14% 15-year EPS growth at 11x, >50% insider plus 20% Mitsui ownership making buybacks 3x more potent; now subject to a take-private offer for the remaining 17%.
The capital-allocation model: tripled acreage to 18M acres while doubling acres per share (~6%/yr), no need to reinvest cash flow in new royalties, and executives required to buy 2x–5x salary in stock with cash.
The 30-year anchor holding — West Texas surface & royalty acreage as a perpetuity, with buybacks compounding acres per share; four 40–53% drawdowns in eight years and still up 4x.
One of the few recent IPO participations alongside LandBridge — Permian water midstream fitting the same hard-asset, royalty-like selection criteria; no ticker or separate thesis given in the commentary.
The other recent IPO participation named — cited only as fitting the firm's hard-asset, valuation-driven criteria; no ticker or financials given in the commentary.
Private prediction market blamed for the exchange selloff — but incumbents can list the same event contracts, so it reads as proof of demand, not displacement; its 6x bitcoin perpetual is a wager, not a hedge.
Named only in the bitcoin adoption curve as an early institutional holder that recognized the fiat-debasement hedge; a network datapoint, not a recommendation.
Cited only as evidence tokenization is under way across the regulated exchanges (SEC-approved token-form trading and DTCC settlement); no valuation view offered.
ICE owns 17% ($1.6B) for the crowd-sourced event pricing and volume data feeding its information business — not the wagering, which the commentary says serves no real-economy need.
The compounding "snowball" exhibit turned caution — AI data-center capex has pushed it from free-cash-flow positive to negative with large-scale borrowing, at 36x, inside the hyperscaler group Tjernstrom's May letter effectively told holders to sell.
Same sale rationale — the 2015 setup (gold below cash cost, capital-starved miners signing double-digit-rate royalties) is gone, and today's $4,100 gold is embedded as a premium rather than a discount.
Named in the same index-concentration critique and among the AI data-center hyperscalers Fredrik Tjernstrom's May fundholder letter was "essentially a sell recommendation" on.
The decentralized venue behind the May 15 exchange selloff — but a perpetual has no expiry and no delivery, so it can't hedge a commercial exposure; leveraged retail gamification, not institutional competition.
Named in the 49%-Information-Technology index-weight challenge and inside the Mag 7 priced at 150x run-rate free cash flow with $108B of stock comp not added back.
The third name in the index-weight challenge — 49% of the S&P in Information Technology presented as a capital misallocation, and the "cheap" 24.8x Mag 7 P/E dismissed once free cash flow is used instead.
Reduced/sold on valuation, not on gold — ~2x NAV against consensus NAVs discounted at 3–5%; a fade to 1.5x NAV over five years cuts a 15% compounder to an 8.6% return.
One line: Own the rare business models that can compound for decades — land, royalties, securities exchanges, bitcoin — bought on business analysis rather than commodity or macro forecasts, and held long enough for the geometric part of the curve to arrive; sell them when valuation, not quality, breaks the arithmetic.
Time is the scarce input, not insight. Value compounds geometrically while time passes arithmetically, so most of the return lands in the final stretch of the holding period — the water glass is half full one time step from the end; the snowball has 15% of its mass at half the clock. Watching price instead of elapsed time is what makes investors sell early. (2026-JUL-31)
Asset-light perpetuities. Land is the longest-lived asset, always shrinking in acres per person, repurposable to higher uses, and not even a sector in the equity indexes; royalty contracts run 20 years or life-of-mine; national exchanges have operated since the 1800s. The metric that matters is units of the perpetuity per share — acres per share, not total acreage. TPL · LB · PSK.TO
Securities exchanges are the misread franchise. Near-zero incremental cost per trade turns 6% revenue growth into double-digit EPS growth; crises and technological shifts have been business opportunities, not displacement, for two centuries. The 2026 perpetual-futures and prediction-market scare left ICE and CME at their cheapest valuations since 2009, CBOE the cheapest in three to five years, and MIAX growing revenue 40%. A perpetual with no expiry and no delivery cannot hedge a commercial exposure — it competes for gamblers, not institutions.
Valuation ends positions, quality doesn't sustain them. The precious-metals royalty companies were bought in 2015 because gold below cash cost let them sign double-digit-rate contracts — not because of a gold forecast. At ~2x NAV built on 3–5% discount rates, a fade to 1.5x NAV over five years turns a 15% earnings compounder into an 8.6% return, so WPM and FNV were reduced or sold. (2026-JUL-31)
Bitcoin: same thesis as 2015, and "right on schedule." Its supply schedule is programmed, so the cost of production doubles every four years (~19%/yr) and acts as a price floor. April 2028 halving math: ~$150k–$250k from production economics, $270,438 from a Metcalfe-derived sub-t^6 power law. Expect ~33%/yr rather than the 2016 era's 90% — each 10x takes longer than the last.
Indexation is a tool, and its distortions are the opportunity. Not hostile to Bogle's original idea, but ETFs became the marginal price-setter of what they already own, leaving the S&P at 49% Information Technology. The excluded class — sub-index market caps, buyback-shrunk floats, insider-heavy ownership, "wrong-shaped" conglomerates — carries discounts imposed for non-economic reasons (the ETF Divide) or for value realization too distant for a relative-return industry (the Equity Yield Curve). AN · PAG
Bearish the index's core. The Mag 7's quoted 24.8x P/E becomes 150x run-rate free cash flow once negative-FCF members are counted generously and $108 billion of stock compensation is left in as an expense; Fredrik Tjernstrom's May fundholder letter on the AI data-center hyperscalers was "essentially a sell recommendation," and even AMZN — the firm's own compounding exhibit — has crossed from free-cash-flow positive to negative on data-center capex. (2026-JUL-31)
Labels are irrelevant; substance isn't. The firm criticizes the ETF industry while issuing active ETFs, and calls most IPOs engineered to be sold to the public (56% below issue after three years) while anchoring LandBridge's. Each decision is a business-and-valuation decision made "definition-agnostic."
The product
What it is: Horizon Kinetics LLC is an investment adviser — separately managed accounts, private funds and a set of actively managed ETFs — whose public face is this quarterly Market Commentary plus periodic client roundtable webinars. Grounded in what the 2026-JUL-31 commentary says about itself; revise as later commentaries reveal more.
Offering
What it is
How the firm runs it
Seen in the index
Separately managed accounts
Core Value and other strategies run in a client's own account, with a relationship manager
The default structure; the commentary tells clients to contact their relationship manager for the new Spin-Off strategy positions
TPL, MIAX — disclosed as significant holdings across advisory accounts
Inflation Beneficiaries ETF (INFL)
Active ETF of hard-asset, royalty-like and inflation-benefiting businesses, managed by James Davolos
Positioned as a "completion fund": its S&P-overlapping holdings are a 0.57% index weight, so it supplies exposure the index lacks
INFL
Blockchain Development ETF
Active ETF established off the tokenization/next-generation-exchange thesis
"The existence of crypto-blockchain-based companies like Kalshi or Polymarket aren't a threat to incumbent exchanges so much as an indicator of… expansion possibilities. Which is why we established the Blockchain Development ETF."
ICE, CME, CBOE, NDAQ, MIAX
Japan Owner Operator ETF
Local Japanese owner-operator businesses beneath the mega-cap index umbrella
Built because ETF exposure to Japan is limited to globally-revenued mega-caps with little exposure to the local economy
—
Japan Special Opportunity strategy
Listed captive subsidiaries the Japanese government wants made independent or taken private
Graham-and-Dodd discounts with a policy-driven catalyst; no ETF yet — "in this case, we wish we had an ETF" — so clients open an advisory account holding a few dozen positions individually
—
Quarterly Market Commentary
Free, public PDF (41 pages this quarter), self-limited to "a maximum of 90 minutes" of reading
Structured around actual client questions, with the reasoning and arithmetic shown rather than asserted — explicitly to avoid "planted axioms"
Every row on this hub
Roundtable webinars & The Devil's Advocate
Topic-specific client webinars; a short-sale research report written by analyst Fredrik Tjernstrom
A recent roundtable covered "the rapidly deteriorating business model of the five AI data-center hyperscalers" with Tjernstrom and James Davolos; Tjernstrom's May fundholder letter was "essentially a sell recommendation" on them
AMZN, GOOGL, META, NVDA
How it serves retail investors
The commentary is free and shows its work. Valuation claims are rebuilt from cash flows in public — the Mag 7 at 150x free cash flow, the royalty companies' 8.6% compression math, bitcoin's production-cost model — so the method is reusable even by someone who will never be a client.
The ETFs are the retail-accessible wrapper. The firm's stated reason for launching them is that some clients prefer or require an ETF, or cannot open a standalone account — INFL is the one with a substantive public thesis in this commentary.
It answers uncomfortable questions on the record. This quarter's agenda came from clients challenging apparent inconsistencies (why issue ETFs, why buy IPOs) and from an alarm about a holding's price decline — answered with the drawdown history rather than reassurance.
Caveat: the material is written for existing clients. Some positions (the Spin-Off strategy names, the unnamed Asian airport and its parent) are deliberately withheld and routed to a relationship manager, and disclosed holdings like TPL and MIAX come with affiliate-status trading restrictions that don't apply to an outside investor.