Tornado.cash: Why the Name Outlived the Interface

Prepared by the editorial team. Updated August 31, 2026.

Tornado.cash is searched for far more often than the thing it once named still exists in the form people are picturing. A brand name acquires its own momentum, and that momentum outlasts the interface, the coverage and the legal position that produced it. The result is an information environment where the most visible answers are frequently the oldest ones.

Why does a name keep drawing attention after the thing it named changed?

Because a name is learned once and reused indefinitely. People search for what they remember hearing, and memory is formed by a burst of coverage rather than by a subsequent quiet correction. Demand for a name therefore decays slowly, while the underlying facts can change completely in a single week.

This case produced an unusually large formative burst. The August 2022 sanctions action generated worldwide coverage across mainstream news, financial press and technology media within days, and that coverage established the association most readers still carry. Later developments were reported in narrower outlets to a much smaller audience, so the correction never reached the population that formed the impression.

The mechanics of search reinforce the gap. Query volume signals to publishers that a topic is worth writing about, and the resulting articles are written for the query rather than in response to a new event. A name with residual demand accumulates content indefinitely, much of it recycled, and the recycling reproduces whatever the dominant framing was at the time the topic became popular.

What kinds of results fill the gap this leaves?

Three kinds dominate: articles that were accurate when published and were never revised, summaries that recycle those articles without checking them, and pages that occupy the familiar name in order to trade on recognition. All three benefit from the same thing, which is a reader who assumes a name reliably identifies a system.

Stale-but-once-accurate material is the largest category and the least malicious. A piece written in September 2022 describing the name as designated was correct on the day it was published, and nothing about it signals to a later reader that its central factual claim has since changed. Undated pages are the worst offenders, because they remove the one cue that would let a reader discount them.

Derived summaries compound the problem rather than correcting it. When new material is assembled largely from existing material, the proportions of the original corpus carry through, and a claim repeated across hundreds of sources reads as consensus regardless of whether any of those sources checked it. This is why the same superseded statements resurface in content published years after they stopped being true.

The third category differs in kind. Recognition has value, so pages appear that use a well-known name to imply a connection they do not have, sometimes soliciting information from visitors who believe they have arrived somewhere official. A name is a pointer that can be re-pointed, and treating it as an identity is what makes the substitution work.

How can you date an undated claim you find about it?

You date it by looking for explicit publication and revision dates, inferring a floor from the last event the text mentions, comparing against an independent archive snapshot, tracing each claim to a primary document, and recording what was true on the day you checked. This is a verification method, not guidance about any service.

Step 1: Look for an explicit publication and revision date

Check the visible page and the underlying markup for a stated publication date and any separate revision date, because a page carrying only one of the two may have been quietly edited or may never have been reviewed. Structured data in the page source often preserves a date the layout omits.

Step 2: Date the text by the last event it mentions

Identify the most recent event the text refers to and treat that as the earliest possible writing date, since a piece that stops at a known milestone was almost certainly written around it. Text that ends at the 2022 designation and never mentions later rulings dates itself.

Step 3: Compare against an independent archive snapshot

Look the page up in a public web archive to see when it first appeared and whether its wording has changed, because an archive records history that the page itself has no obligation to disclose. A first capture long after the events described is informative on its own.

Step 4: Trace each factual claim to a primary document

Take the specific claims that matter to you and locate the agency notice, court filing or official record behind each one, because a primary document carries its own date and scope. Most contested statements dissolve at this step into something narrower.

Step 5: Note what was true on the day you checked

Write down the conclusion you reached and the date you reached it, because a status that changes over time cannot be recorded accurately without the date attached to it. An undated conclusion becomes another stale claim within months.

Why do stale results outrank corrected ones?

Because ranking rewards accumulated authority, and accumulation takes time. An article published in 2022 has years of links, references and traffic behind it, while a 2025 correction starts from nothing. Ranking systems measure how established a page is, which correlates with age rather than with current accuracy.

Publishers rarely revisit old pages either. Updating an article that already performs well carries editorial cost and no obvious benefit, so the incentive is to publish something new instead. The old page keeps its position, the new page competes for a different query, and the superseded claim stays in circulation with nothing pushing against it.

The older framing is also easier to write about. A designation is concrete and dramatic, while the later record is a series of qualifications: a ruling limited to statutory scope, a delisting that is not exoneration, a verdict split across counts. Nuance compresses badly, so the version that circulates is usually the earlier one.

Signals that date or discredit a result

A few observable features reliably distinguish current material from material that has merely survived and accumulated authority. The table below lists the most common signals, what each one suggests about a result, and what it cannot establish, offered as general research guidance rather than as an assessment of any particular page.

Signal	What it suggests	What it does not establish
No visible publication date	The publisher is not accountable for currency	That the content is wrong on every point
Text describes the name as presently sanctioned	Written before March 2025 or never revised	That the author intended to mislead anyone
No citation to any primary document	Claims are inherited from other secondary writing	That a primary source does not exist
Archive shows a first capture only recently	The page is newer than the events it narrates	Who operates it or for what purpose
Visitors are asked for a wallet recovery phrase	A strong indicator of a fraudulent page	Nothing further; no legitimate reason exists to ask

None of these signals is decisive alone, and the last one is a category apart from the others. Two or three appearing together are usually enough to set a result aside and look for a primary source instead.

What does the current record actually say?

It says the position changed repeatedly and is not settled. The name was designated in August 2022, a federal appeals court held in November 2024 that the immutable contracts could not be designated, and the name was removed from the sanctions list in March 2025, so it is not currently designated. Criminal proceedings continued separately.

Each of those steps is narrower than its usual summary. Van Loon v. Department of the Treasury decided a question about the scope of sanctions authority rather than the legality of any conduct. The March 2025 removal from the sanctions list is not an exoneration and has no effect on criminal statutes covering money laundering or unlicensed money transmission.

The individual cases remain live and are the reason the subject continues to generate coverage. Alexey Pertsev was convicted in the Netherlands in May 2024 and sentenced to 64 months, then conditionally released in February 2025 with an appeal pending. A jury convicted Roman Storm in August 2025 on one count of conspiracy to operate an unlicensed money transmitting business and deadlocked on two others, and a retrial is scheduled for April 26, 2027 before Judge Katherine Polk Failla in the Southern District of New York.

Figures repeated in coverage deserve the same scrutiny as legal conclusions. Treasury asserted in 2022 that more than seven billion dollars had been laundered through the service since 2019, a figure researchers have disputed, and that assertion has been reproduced ever since as though it were established. Attribution matters, and a number carried from a press release into a thousand articles gains circulation rather than confirmation.

Frequently asked questions

Does a high position in search results indicate a page is legitimate?

No. Ranking reflects signals such as links, engagement and topical relevance, none of which measure whether the operator of a page is who it appears to be. Paid placements and newly created pages can also appear above long-established sources.

Why do automated summaries sometimes give outdated answers about this?

Summarizing systems draw on large volumes of published text, and most of the text about this subject was written between 2022 and early 2025. When the corpus is dominated by one period, a summary tends to reflect that period unless it is explicitly grounded in a current primary source.

Is an archived snapshot evidence of what a site shows today?

No. An archive records what existed at the moment it was captured, which is exactly what makes it useful for dating a claim. It says nothing about who controls the same address now or what a visitor would encounter there.

Does a valid security certificate mean a site is trustworthy?

No. A certificate confirms that traffic to a name is encrypted and that the holder demonstrated control of that name, and such certificates are freely available. Encryption is a property of the connection rather than an assurance about the operator behind it.