Decide which of four things you mean first. “Protect” covers theft, your own errors, creditors and claims, and losing access when you die or cannot act. Those need different tools, different professionals, and they work against each other. In my experience most people arrive worried about one of the four and buy the arrangement built for a different one.
Part of our guide: Family Office.
The short version
- Theft, error, claims, and access loss are four unrelated problems. Naming which one you mean is most of the decision.
- Hardening against theft usually weakens survivability. The arrangement an attacker cannot reach is often the one your executor cannot reach either.
- Legal protection is real and bounded. A transfer aimed at an existing creditor can be unwound, and Wyoming gives that creditor two years to bring the claim (W.S. 34-14-210).
- The largest measured channel of loss is persuasion. The FBI logged 181,565 complaints with a cryptocurrency nexus in 2025 and $11.366 billion in reported losses.
- None of this touches market risk. That is a concentration question with its own answer.
Four questions wearing one word
The four versions sound identical when spoken and resolve differently in every respect. Sort yours before spending money.
| What you actually mean | What takes the assets | Where the work sits |
|---|---|---|
| Theft | An attacker, an insider, a convincing stranger | Keys, devices, signing procedure |
| My own error | You, on an irreversible action | Limits, a second signer, verification |
| Creditors and claims | A judgment, a divorce, a business dispute | Title, timing, state law |
| Loss of access | Nobody at all | Documents, named people, reachable recovery material |
Notice the last row. In the first three somebody ends up holding the assets, so there is a counterparty and sometimes a recovery. In the fourth the coins sit at an address forever, visible to everybody and spendable by no one, with nobody to appeal to. That asymmetry is why I put access loss at the top whenever the honest answer to “who else can reach this” is nobody.
Why the four pull against each other
Stacking every recommendation you read produces a worse position than picking two deliberately, because the limbs conflict.
Theft controls fight access. Every step that narrows who can move assets narrows who can recover them. A passphrase held in one head, a seed split across places nobody mapped, a quorum tuned so no plausible group of attackers can meet it: each answers theft, and each raises the odds the assets become unreachable by anyone. Custody is that trade compressed into one decision.
Legal controls fight operations. An entity or a trust adds people, documents, and steps, and each addition is a place an error can happen. A structure nobody can operate under pressure has swapped one exposure for two.
Succession fights secrecy. A plan somebody can follow is a plan somebody can read, so every workable version ends in a rule about who knows what and when.
What the legal limb does, and where it stops
Entity and trust work can improve creditor posture. It reduces certain risks and does not eliminate them, and the failure modes are worth knowing before you pay.
The first limit is timing. Almost every state has adopted some form of the uniform fraudulent transfer rules, under that name or the later name Voidable Transactions Act (Uniform Law Commission). They let a creditor attack a transfer after it happens, and Wyoming, which still runs the earlier text, lists the factors a court weighs on intent:
“(b) In determining actual intent under paragraph (a)(i) of this section, consideration may be given, among other factors, to whether: (i) The transfer or obligation was to an insider; (ii) The debtor retained possession or control of the property transferred after the transfer; … (iv) Before the transfer was made or obligation was incurred, the debtor had been sued or threatened with suit; (v) The transfer was of substantially all the debtor’s assets”
Wyo. Stat. Ann. § 34-14-205(b)
Read that against a hurried crypto entity. Sole member, so the transfer went to an insider. Same seed phrase and same wallet afterward, so the debtor kept control. Everything moved at once, so it was substantially all the assets. Formed the month a demand letter arrived. Four listed factors from one weekend of work. No single factor decides anything, and the statute says as much, but four together is an argument handed to the other side.
The second limit is that the clock runs from the transfer rather than from the claim. Wyoming extinguishes the claim two years after the transfer, or six months after it could reasonably have been discovered if that falls later (W.S. 34-14-210). Federal bankruptcy law carries its own two-year reach-back for a transfer made “with actual intent to hinder, delay, or defraud any entity” (11 U.S.C. § 548).
The third limit is that none of it addresses your own conduct. A charging order under W.S. 17-29-503 governs what a judgment creditor reaches through a membership interest, and says nothing about a claim brought against you personally. The Wyoming LLC material covers the mechanics. Structure earns its keep when it is built early, funded the day it is formed, and operated as a real thing.
The loss almost nobody structures for
The measured picture does not match the one people plan against. The FBI’s Internet Crime Complaint Center puts it plainly:
“Cryptocurrency investment fraud was the highest source of financial losses to Americans in 2025 with $7.2 billion reported in losses.”
FBI, 2025 IC3 Annual Report
Across all crime types, complaints with a cryptocurrency nexus reached 181,565 in 2025, with $11.366 billion in reported losses and 18,589 complainants reporting more than $100,000 each (2025 IC3 Annual Report). Self-selected complaints make that a floor.
The shape matters more than the totals. Investment fraud, impersonation, and extortion dominate, which means the holder authorized the transfer. A hardware wallet signs that happily, and so does a trust, an LLC, or a custodian following valid instructions. The controls that work here are procedural: a written rule about whose instructions you act on, a delay before the first transfer to any new address, and a second human above a stated threshold. NIST’s key management guidance covers the cryptographic half.
What I actually see
The most common thing is a tool bought for the wrong limb. Somebody reads about entity formation after a scare with a phishing message and forms a company. The company is fine and the phishing exposure is untouched, because the wallet, the device, and the habits never changed.
The second is a security posture that turned into a succession failure while everyone congratulated themselves. The strongest cold-storage arrangement I have looked at had material split across two states, a passphrase living in one person’s memory, and no inventory. Against theft it was excellent. Against the death of one person it was a total loss, and nobody involved had connected the two facts.
The third is legal work done at the wrong moment. Structure built while nothing is on the horizon is planning. The same structure built the month a claim appears reads worse for every statutory factor it touches.
The check I would run takes an hour. List every wallet, account, and device, then answer four questions about each in writing. Who besides me could move this today, and how would I find out? What is the largest irreversible action I can take with nobody else involved? If a judgment landed tomorrow, who holds title, and when was it transferred? If I were unreachable for ninety days, who reaches it, using what, and have they ever done it? Whichever question takes longest to answer names the limb to work on. For most people it is the fourth.
Where this goes wrong
The failures cluster where two limbs meet and each professional assumed the other one had it covered.
The specific ones: a signing quorum no group of attackers can meet and no group of survivors can either; recovery material split three ways with no map to the three; an entity formed with a demand letter already on the desk, still signing from the same personal wallet; a trustee holding a hardware wallet the trust instrument never authorized, which is a live question with a real answer; records good enough for a tax return and useless as proof of who owned what, which is the usual pattern; and the whole position moved to a “safer” address on the instruction of somebody impersonating support.
The decision rule
Work in this order, which is deliberately not where people start.
- Name the threat before naming the tool. One written sentence saying which of the four you mean.
- Rank all four against your own facts. A single holder with no dependents and the owner of an exposed business rank them in opposite orders.
- Close access loss first wherever it is open, because nobody can repair that failure afterward.
- Do the legal work while it is still planning, before any claim is foreseeable, and fund the structure the month you form it.
- Keep the theft controls and the succession controls on one page, so tightening either shows what it just broke.
- Write down whose instructions you act on, how you verify them, and what delay applies to a first transfer to a new address.
- Make the records prove title, a higher standard than a tax return needs, because a court, a custodian, and an heir all read them for that.
- Recheck on events rather than dates: a new device, a new person, a new custodian, a claim, a move.
Where this sits
This question sits above the cluster rather than inside it. Custody answers the theft and access limbs. Trusts and estate documents answer what happens when you cannot act, and the common estate errors are worth reading before anything gets drafted. Wyoming LLCs answer the claims limb, and custody inside an entity is where the entity and the key meet. Start with whichever of the four questions you answered slowest.
An attorney drafts the trust, a CPA files the returns, a custodian holds the keys, and none of the three sees the other two’s work or is paid to look for it. The arrangement is only as good as whoever reconciles the pieces, and in most families that job belongs to nobody. Give it to a named person, yourself included, and put a date on the next review.
Sources
- FBI, 2025 IC3 Annual Report (Internet Crime Complaint Center)
- Wyoming Uniform Fraudulent Transfer Act, Wyo. Stat. Ann. §§ 34-14-205, 34-14-210 (Wyoming Legislature, Title 34)
- Wyoming Limited Liability Company Act, Wyo. Stat. Ann. § 17-29-503 (Wyoming Legislature, Title 17)
- 11 U.S.C. § 548, fraudulent transfers and obligations (Cornell Legal Information Institute)
- Uniform Law Commission, Voidable Transactions Act
- NIST, Special Publication 800-57 Part 1 Revision 5, key management
- IRS, Digital assets
Related
- What is crypto wealth management?
- Crypto concentration risk management
- Crypto custody for LLCs
- Common crypto estate planning mistakes
- Crypto wealth planning for Bitcoin millionaires
- Crypto wealth management
Last updated: 3 August 2026.
This article is general education, not legal, tax, or investment advice. Custody, entity, and trust arrangements can reduce certain risks but do not eliminate them, and a transfer made to defeat a creditor who already exists can be unwound by a court. Talk to a qualified attorney and CPA about your own situation.
