Holding company structures, and when a founder is too small for one

Holding company questions usually arrive in the same shape: the business is working, someone mentioned asset protection or an exit, and the founder now suspects the single company they have is naive.

It usually isn't. A holding structure is a real answer to a small number of real problems, and an expensive answer to everything else. The cost is not the formation fee — it is that every ongoing obligation you have now happens twice, plus a set of new ones that only exist because there are two entities.

What a holdco actually does

Stripped of the diagrams, a holding company does exactly one thing: it separates ownership of an asset from the activity that carries risk. Everything else follows from that.

• Ring-fencing. A claim against the operating company reaches the operating company's assets. Anything held above it is a step removed.
• Separating multiple businesses so one failing does not take the others with it.
• Holding intellectual property apart from operations, licensed down for a royalty.
• Providing a single vehicle for several founders or investors to own several things through.
• Exit shaping. Some jurisdictions exempt gains on the sale of a qualifying subsidiary from tax at the holding company level.

Note what is absent. It does not reduce the tax on operating profit. It does not protect you from your own conduct. And it does not make a business more credible — if anything, layered ownership makes banking and customer due diligence harder, not easier.

The running cost nobody quotes

The six cases that justify one

1. Genuinely separate businesses with separate risk

Not product lines — businesses. Different customers, different liability profiles, plausibly sold to different buyers. A software product and a physical-goods brand under one owner is a real case. Two SaaS products sharing a codebase and a team is not.

2. Intellectual property worth separating

The IP must be substantial, genuinely capable of being licensed, and able to bear a royalty that would survive a transfer pricing review. Moving a codebase into a second entity and charging a number you invented is worse than doing nothing: it creates a taxable flow, a documentation obligation and an obvious question, in exchange for protection you could have got from insurance.

3. Real estate

The oldest and least controversial use. Property carries long-tail liability and sits still; operations carry short-tail liability and move. Keeping them apart is standard practice and needs no cleverness.

4. Multiple owners across jurisdictions

Where founders and investors are in several countries, a common vehicle in a neutral, well-understood jurisdiction is genuinely simpler than everyone owning slices of everything directly. The structure exists to make ownership administrable, which is a commercial purpose that survives scrutiny.

5. A planned exit into a participation exemption

Some jurisdictions exempt gains on the disposal of a qualifying shareholding. Where one genuinely applies to you, holding the operating company beneath such an entity can be decisive at sale. Three conditions attach: the exemption must actually apply on your facts, the holding period requirements must be met, and the structure must have been in place long enough that it does not look assembled for the transaction. That last one is what the principal purpose test is for.

6. You need a US topco for US investors

The most common legitimate case for readers of this blog, and the subject of the next section.

The Delaware flip

A founder outside the US builds a company in their home country. US investors want to invest, and US investors overwhelmingly want to buy preferred stock in a Delaware C corporation — their funds' documents frequently require it, and the tax treatment of a foreign holding is unattractive to them.

The flip inverts the structure: a new Delaware corporation is formed, the existing shareholders exchange their shares in the home-country company for shares in the Delaware entity in the same proportions, and the original company becomes a wholly-owned subsidiary. Operations continue where they are.

It works. It has one hard rule.

Two secondary points. The flip creates a permanent US filing footprint that does not go away if the round does not happen. And it does not move your operations — the home-country subsidiary keeps its employees, its obligations and its own tax residence.

The cheaper alternatives

The too-small test

All of the following true means you are too small, and the structure will cost you more than it returns.

1. 1One owner, or a small group who all live in the same country.
2. 2One business, with one revenue model.
3. 3No IP that a third party would pay a defensible royalty to use.
4. 4No outside investors, and none being sought in the next twelve months.
5. 5No property.
6. 6The realistic worst-case claim is covered, or coverable, by insurance.

A rough financial framing: a two-entity structure with adequate substance is realistically low thousands a year before advice, and considerably more once any part of it is outside your country of residence. It should be saving or protecting a multiple of that, reliably, before it is worth doing. If the benefit is speculative and the cost is certain, the trade is bad — and it is bad every year, not once.

The honest version: most founders asking about holding companies have been sold the idea by someone who forms holding companies. The question worth asking instead is which specific asset is being protected, from which specific claim, and what the same money would buy in cover. That question answers itself surprisingly often.