Open in interactive viewer → charts, metric popovers & call review

Horizon Kinetics: $36 Million of Cash, a $25 Million Buyback, and a Stock That Won't Participate

Q3 was flat, the loss was mostly an accounting allocation — and management spent the hour arguing the market pays nothing for a $10 billion AUM franchise sitting on scarce hard assets.
SLGD · Earnings Call · 2026-11-17

A flat quarter, an ugly loss line, and a $0.8 million office move

Strip out the noise and Horizon Kinetics' third quarter was remarkably ordinary: “advisory revenues of $17.8 million, which is essentially unchanged or flat as compared to 2025's third quarter, but it is down 5.3% sequentially” — Mark Herndon, Chief Financial Officer · 2026-11-17. Operating income fell to $1.8 million, down 40% year over year, and the reason was mundane rather than cyclical — a right-of-use impairment plus duplicative rent from relocating the New York office, a cost management had pre-announced last quarter. Operating expenses rose $1.3 million to $16.3 million. The headline number looks catastrophic and is almost entirely optical: a $59 million quarterly net loss, of which $41 million was allocated to the redeemable noncontrolling interest inside the consolidated investment products, leaving $18.4 million — $0.98 a share — attributable to shareholders. That is a mark-to-market event, not a cash event. Assets under management were $10 billion at September 30, up from $9.6 billion at year-end 2025 but down from $10.8 billion in Q2, and the swing is almost a single security: Texas Pacific Land fell 25% in the quarter (still up 13% year to date) while the bitcoin-linked holdings ran up 43% in the quarter and remain down 6% year to date. For the nine months, GAAP net income was $35.8 million, $1.92 per share, flattered by an $18.1 million first-quarter performance fee tied to expiring trading restrictions on MIAX. On the supplemental adviser-only basis — which pulls incentive fees back into the revenue line where they economically belong — management advisory fees were $77.9 million against $61.2 million, up 27%. The balance sheet is the anomaly for a company this small: $36.4 million of cash, an $86 million investment portfolio, $12 million of digital assets, roughly $243 million in private funds, no third-party debt, and long-term liabilities confined to office leases. The board added a $0.13 quarterly dividend (a trailing $0.507 annualized) and, more pointedly, authorized a $25 million repurchase program.

The pitch: own the choke point, not the trade

Steve Bregman used most of his airtime on a macro argument, and it is genuinely differentiated from the standard asset-manager growth story. The thesis: sovereign deficits plus commodity depletion plus electrification mean the binding constraint on AI is not chips but hard assets — natural gas, water, and land in the right place.

The choke point, the limiting factor for AI development for electrification of the economy is there's a need for hard assets, hard assets like natural gas, like water, like land, meaning land located in a specific place that's also suitable for a data center, far from agrarian land or population centers is really in fairly short supply.

Steven Bregman, Executive or Senior Management · 2026-11-17
What distinguishes this from ordinary commodity cheerleading is the second derivative: they claim not to need the data center to show up. “They're profitable on their own, and we believe they're undervalued. And any such further development is simply pure optionality.” — Steven Bregman, Executive or Senior Management · 2026-11-17 The supporting facts he offers are checkable: Texas Pacific Land at roughly a $20 billion market cap is just 0.03% of the S&P 500; the total market cap of natural gas royalties in the U.S. is something like $600 million; the energy sector is a little over 3% of the index versus its historical weight. Peter Doyle's version is blunter — they built what he calls a horizontal monopoly across West Texas, owning "the toll roads" into the powered land that power-hungry operators will need.

The tape is voting the other way — for now

Here is the tension the call cannot resolve. If the thesis is that compute scarcity pulls hard assets higher, the last 90 days have been an outright repudiation of the first half of that chain. The market's data center capacity complex is down roughly 29% over three months across ten names, and the semiconductor and compute-infrastructure cohort has been hit harder still — High Performance Computing has shed nearly 30% across twenty tickers in the same window. The whole AI-buildout trade has de-rated. Horizon is not immune, because it is not neutral on the collateral: TPL, the single largest driver of its AUM, gave back a quarter of its value in Q3. The argument that the assets are profitable without the data center is intellectually sound and completely untested against a tape that marks the whole chain down together, correlations converging to one exactly as Bregman half-joked. The second leg of the book has the same problem. The price of Bitcoin is down roughly a third over 90 days across the crypto complex, while Horizon's digital assets carried at $12 million and the bitcoin-linked securities inside its funds were up 43% in the quarter alone. Volatility, not direction, is the reliable feature — and that is precisely what makes the GAAP earnings line un-analyzable quarter to quarter.

Green shoots, ReFlow, and buying back the discount

Operationally the most consequential item was barely a paragraph: joining the ReFlow program, which lets the mutual funds deliver securities in kind instead of selling low-basis positions on redemptions. “ReFlow allows us to access market liquidity without having to sell individual securities. And that puts the mutual funds now on much more equal footing with ETFs.” — Peter Doyle, Executive or Senior Management · 2026-11-17 Doyle estimates the coming Paradigm Fund capital gain would have been about one-tenth the size had the structure been in place for 2026. For a firm whose mutual fund lineup has been bleeding revenue to ETF substitution — the Inflation Beneficiaries ETF grew revenue 26% in the quarter and 30% year to date, but was "predominantly offset" by mutual fund declines — that is the closest thing to a structural fix on the call. INFL holds roughly $1.5–1.6 billion, and the first real marketing campaign begins "literally this week." Then there is the buyback, which is really a discount to NAV trade on their own equity. “From where we stand, the marketplace right now is placing no value on the asset management firm.” — Peter Doyle, Executive or Senior Management · 2026-11-17 Twenty-five million dollars against a stock that trades a few thousand shares a day, with no 10b5-1 plan in place — so execution will depend on bloc availability, and management conceded as much. The disclosed run-rate is about $80 million of revenue. Two caveats deserve equal billing. First, the incentive-fee pipeline is thin: about $6.9 million of unearned fees as of September 30, of which $6.7 million sits behind liquidity restrictions similar to MIAX — against a company record north of $50 million booked in Q4 2024. Second, when a shareholder asked about Fermi and the clarity supposedly arriving after the election, Doyle deferred — recent elections are not, on this call, something management will talk about at the security level. So: a sub-$1 million reported market cap in our metadata is plainly a data artifact, but it rhymes with the story being told. Either Horizon owns a scarce-asset book that the index-constrained market structurally cannot price — in which case the buyback and the dividend are the right capital allocation and the $0.98 quarterly loss is noise — or the fee stream is quietly shrinking (advisory fees down 1.2% for nine months) and the choke-point narrative is doing the work that revenue growth is not. This quarter offered real evidence for both readings and no tiebreaker.