Closed-ladder markets: an investment thesis.

The name of the firm is the thesis of the firm. A closed-ladder market is a market in which the regulatory, cultural, or technical complexity creates a moat for whoever climbs the ladder first, and in which most venture dollars can't or won't follow. The won't is doing more work in that sentence than the can't. We have spent the last six years operating in markets that look closed from the outside, and the consistent observation is that the moat is built less by what an incumbent did and more by what the next entrant decided not to do. The ladder closes behind the climber because nobody else picked the slow, sensitive, or unsexy route up. We did. This is the formal write-up of why.

What is a closed-ladder market.

A closed-ladder market has three properties. Take any one of them away and the thesis stops applying. All three must be present at the same time.

One. The market has structural complexity that takes a long time to learn. Regulatory law, institutional procurement cycles, consent regimes involving multiple institutions, vendor-approval lists, professional-licensing gates, hardware-validated supply chains. The complexity is not solvable by hiring smart engineers; it is solvable only by the calendar.

Two. The market is small enough that the standard venture-fund math does not work on it. A fund with a $300M vehicle needs every position to be capable of returning the fund. A market that supports a sustainable $40M ARR business at exit is, for a $300M fund, a position the partner cannot defend in the LP meeting. The market is not too small to be a good business. It is too small to be a good fund position.

Three. The market is sensitive, slow, or unsexy in a way that filters out attention. The cap table is uglier than a SaaS round. The customer logos do not get retweeted. The board of advisors cannot be photographed at a conference. The press cycle is hostile or absent. The category does not produce listicles.

When all three are present, the market is closed-ladder. The dollar that goes in compounds against a much thinner field of competing dollars than the categorical-tailwind dollar going into adtech or AI-agents in the same vintage.

The categories we are in.

Five sectors meet all three criteria in 2026. The list is not exhaustive. It is the list where we have written checks or are actively working a position.

• Regulated controlled-substance supply chains. DEA-Schedule pharmaceutical logistics, the audit substrate underneath them, lab-to-pharmacy chain of custody. The learning curve is measured in years and the procurement cycle is measured in fiscal calendars. The competitive set is six firms. The dollar buys a position the next $300M fund cannot price.
• Reentry technology serving the formerly incarcerated. An app that gets a person home from a state facility into housing, ID, benefits, and a job. Built with a community partner, not for an enterprise buyer. The sensitivity filters out the funds that need the photo op. Linkage is the position in this category.
• Container-image and supply-chain forensics for journalism and counsel. The tooling that lets an investigator confirm that a binary in a leaked container matches a binary in a public artifact registry. Slow. Technical. The customer is a defender, a counsel, or a small newsroom. Three buyers in the country, and the same three buyers come back every quarter.
• Civic data interfaces over hostile public datasets. Humane interfaces sitting on top of state records systems that were designed to discourage public access. The Michigan corrections example — 142,000 people, 520,000 sentences, 83 counties — is one of about thirty in the United States. People Not Numbers is the position.
• Riso, print, and approved-vendor mail into U.S. prisons. The smallest of the five. A market with three vendor channels, a mailroom that is the gatekeeper, and shipping economics that nobody else has bothered to learn. Closed Press is the position.

The five categories share nothing in common at the surface. Pharmaceutical logistics has nothing operationally to do with prison mail. The thesis is what they share underneath: each one looks small, slow, and unattractive to capital that needs velocity, and each one rewards the operator who learned the institutional terrain better than the next operator.

Why this compounds harder than software-eats-the-world did.

The 2010–2020 venture playbook was right about one thing and wrong about a second. The right thing was that software, distributed via the cloud, eats categories that previously required physical infrastructure. The wrong thing was that the dynamic generalizes. It does not. The categories where physical infrastructure still rules — because of regulation, because of consent, because of bonded supply chains — are precisely the categories where the software-eats-the-world play breaks against the rocks of compliance, audit, and licensure. The 2020s have already shown this in healthcare, in education, in autonomy, and in finance.

What compounds harder than software-eats-the-world is the reverse trade: positioning for the market that software does not eat, and then writing the software that becomes the only durable interface to it. Once the interface exists, the next entrant must either license it, replicate the years of compliance work that produced it, or build a worse one and try to underprice it. The first option is the typical exit. The second option takes another decade. The third option does not survive a procurement cycle.

This is the part of the thesis that is most often misread as conservative. It is not conservative. It is the opposite. The conservative play in 2024 was to buy index exposure to AI-agent platforms. The aggressive play is to take a five-year position in a category where the index will not arrive until 2030 because the index cannot do the unsexy paperwork.

How the family of entities works.

The closed-ladder thesis requires more than a fund. It requires a family of entities. Closed Ladder Capital is the for-profit capital arm and writes the equity checks. Closed Ladder Capital Foundation is the 501(c)(3) that funds the boring infrastructure underneath the categories — postage budgets, FOIA filing fees, transcript mailings — because the for-profit capital cannot honestly pay for those without bending. Closed Ladder Advocacy is the 501(c)(4) that does the policy work, because some of the categories require legislative attention that the fund cannot lobby for directly. Closed Ladder Studio is the in-house creative shop that builds the brand surfaces of the portfolio companies, because closed-ladder companies almost never have the budget for an outside agency.

The braid of the family of entities is not for tax optimization. The braid exists because the categories are not legible to any single entity type. A pure fund cannot fund a 501(c)(3) press; a pure foundation cannot take an equity position in a logistics company; a pure 501(c)(4) cannot make grants to operators. The braid lets the network operate on the full surface of the category without distorting any single entity's purpose.

What we underwrite to.

We underwrite to a ten-year hold, with no model assumption that a public-market window opens. The exit is either a strategic acquirer who understands the category, a recapitalization at a fair multiple of a working business, or a planned distribution-in-kind to LPs who want to own the operating asset. The IRR is lower than the LP would see in a hot vintage of a generalist fund. The variance is also lower. The defensibility is structurally different. We are buying durability, not optionality. If you are an LP who needs a 6× in five years, you are the wrong LP. If you are an LP who can sit on a 3× in ten years that is uncorrelated with the rest of your book and is doing public good, you are the right LP. The door is open by letter.

Three claims in that sentence. We expect to be measured against them.