X (formerly Twitter)

TLDR (but please actually read the entire letter here, as its well worth your time): Horizon Kinetics’ Q3 2025 commentary is a love letter to scarcity and tollbooths: own the casinos (exchanges), own the royalty rents (gold/oil/gas/potash), and own the choke points (land and water next to datacenters). If you think the S&P 500 is a diversified portfolio and not an NVIDIA-themed group chat, they would like a word. The Setup: Indexing Was Supposed to Be Boring, Not a Religion Twenty-five years into ETF-ification, the data is well, awkward. Across styles, geographies, and caps, equity ETF annualized returns land mostly in the 7–8% zip code; fixed income ETFs scrape along at ≤3.5%, i.e., negative after taxes and inflation. Meanwhile, Information Technology (now ~46% of the S&P by market value) has turned the index into a momentum maxi—great on the way up, nowhere to hide on the way down. Peak-to-peak since 2000, the “10% forever” promise looks more like a marketing slogan than a law of physics. So Horizon suggests an unfashionable verb: sidestep. Not “underweight” or “tilt.” Leave the index in certain places and rent your exposure from business models that extract value from market activity rather than market narratives. First Pitch: Invest With the House (Exchanges) Securities exchanges are the market’s croupiers. Whether you’re panic-selling a meme coin or hedging soy oil, they clip a spread and smile politely. Over long windows, across geographies (US, Japan, Hong Kong, UK, Singapore). Why? You don’t have to know the winner of “Oil vs. AI vs. Wheat vs. Bitcoin.” Exchanges get paid on —and turnover compounds. (Also, SPY turns over ~100% every . Read that again.) The CPI can tell you 2.9% while your burger combo is screaming 9%. Call it —sectoral pressure that torches specific line items in actual budgets (and P&Ls) while headline indices stay polite. Restaurants see menu inflation and wholesale food cost jumps; investors see a sector ETF (EATZ) that’s gone nowhere fast. Local pain, real money. And the hedge? Think , not operators. In the 1970s oil shock, Exxon didn’t save you; rates rose, capex rose, multiples fell. But an oil passes prices straight through to you without capex/opex drama. Horizon notes delivered a >12x combined return (price + distributions) since 1999 while oil “only” doubled-ish; SBR’s admin costs are about and distributions track the commodity check. Price up? You get paid. No “woulda-coulda-shoulda” capex waterfall. Gold’s up ~145% in three years without CPI fireworks. Why? Scarcity magnifies localized inflation winners. And scarcity cuts two ways: scarce (potash deposits, high-grade iron ore) and scarce (small floats, off-index). Horizon’s example list includes: For a decade, US nat-gas was a shrug: production > demand, price flat-to-down in real terms. Then AI left “the cloud” and demanded an actual . A single modern AI campus can need , i.e., the output of a nuclear plant that takes a decade to build. Batteries cover , not days. Solar requires acreage for 200MW—and you still need a gas plant spinning in reserve. Utilities got re-rated in 2024–2025; gas producers haven’t fully caught the memo. Globally, US LNG exports exploded post-2016; China and India sit on low per-capita gas usage and big AI ambitions. If their power generation creeps toward US per-capita levels, that’s a structural bid for LNG. But gas is also . It moves by pipe, not by wish. Permitting kills projects for sport; storage constraints mean tiny imbalances can smash price. Localized scarcity, meet localized inflation—exactly the kind royalties and bottleneck owners monetize. Thermal power plants (gas, coal, nuclear) are steam engines. Steam needs . The US power sector is one of the country’s largest freshwater users. The “AI campus + gas plant” bundle is therefore also “AI campus + water plan.” In the , you get a trifecta: surplus associated gas (often at the Waha Hub), abundant , and a mature water-handling ecosystem. The result: a quietly massive profit pool—pipelines and disposal wells for water—earning fat operating margins north of per barrel handled. That’s a tollbooth, too. Horizon highlights players like , its now-acquired peer , and (surface rights positioning). Add —distinct business, same logic: own strategic land adjacent to constrained infrastructure and let time + permitting do the multiple expansion. None of these screens into the S&P 500; all of them live where the actually gets built. The thread through exchanges, royalties, and bottlenecks is and . Exchanges scale on code and colocation; royalties scale on contractual skims; water/land bottlenecks scale on permitting, easements, and smallish incremental capex borne by others. You’re not praying for multiple expansion on a cost-of-capital pig that needs new equity every cycle; you’re getting paid while others sweat. Here’s a quick one-page table to keep the categories straight: (We also dropped a live table in your workspace so you can sort/filter.) Just and around the edges and let them grow into bigger weights as the thesis plays out. Some practical scaffolding: You won’t see these at S&P weights until after they’ve re-rated. That’s okay. The whole point is to own the before it gets securitized for the masses. Murray summed up in 1 sentence:

Institutional Insights Library

Horizon Kinetics Just Told You How to Beat the Robot Index—Here’s the Playbook (and the Meme)

Primary Source Material

Institutional Research
Straight to Your Inbox

Institutional Insights Library

Horizon Kinetics Just Told You How to Beat the Robot Index—Here’s the Playbook (and the Meme)

Primary Source Material

Institutional Research Straight to Your Inbox

Institutional Research
Straight to Your Inbox