Memecoin JEANPHIL (Jean Phil): The Man Isn’t Real — Can the Coin Run Again?

On September 17, a video appeared on Instagram. A man in a brown houndstooth suit stands in an apartment — close-cropped blond hair, a curled moustache, a completely flat expression. Then, without warning, he starts shadowboxing at empty air. He doesn’t say a word.

Three days later the same man turns up on a Paris street, on a park bench, outside a café. Same suit, still silent, still punching nothing. One of those clips pulled 5.6 million views and 560,000 likes on TikTok; another shows 15.4 million views on the account page itself. By mid-October the account had posted twelve times and held 244,000 followers on Instagram.

The man doesn’t exist. He is AI-generated.

And on the third day of his fame, someone launched a memecoin in his name. The price made three clean highs, each above the last — September 20, September 25, October 9 — and then went into a slide. It sits at a $3.48 million market cap today, roughly 80% below the peak.

Two things I want to work out. Why can a person who doesn’t exist carry a coin to $16.06 million? And can it do that again?

Why I’m watching a coin that’s already down 80%

Start with what “again” would mean. Not the price back at $0.0166 — that number means nothing on its own. My bar is two things at once: daily volume back above one-fifth of the peak, and a run of addresses that have never held the token before. Price can be moved by a few hundred dollars of orders. Volume and new holders can’t.

On the evening of September 19, at 8:15 in Paris, someone launched a token called Jean Phil on Pump.fun. In the same transaction he bought 13.6 million tokens for himself — 0.4 SOL, about $44 at the day’s rate, 1.4% of supply. Fifty-one minutes later the bonding curve filled and the token opened on PumpSwap. Day one took it from $0.0000445 to $0.00519, closing at $0.0034. Day two peaked intraday at $0.0116 — a market cap near $11.2 million. Day three took it down to $0.00135, a $1.31 million cap: an 88% drop from the September 20 high, inside two days.

That is where an ordinary chart ends. This one kept going.

Between September 22 and 25 it climbed from $0.00164 back to $0.01144 and touched $11.1 million a second time, with a 20% dip on September 23 along the way. On October 2 it was knocked down to $0.00156, then from October 3 it ran a third time and hit $0.01656 intraday on October 9. That is the real top: about $16.06 million.

Three highs: the second came in almost level with the first, the third about 40% above both. Then the slide, down to $0.00238 on October 24 and a $2.31 million cap. That is the low of this decline, not the all-time low — on September 21 it traded at $0.00135, a $1.31 million cap. It’s $0.00358 now: about 51% above the October 24 low, 2.7 times the September 21 low, and 78% below the peak.

The strangest stretch is October 7. That day the BBC ran a fact-check about him — an influencer who doesn’t exist, being used to promote high-risk crypto assets. Dexerto and L’ADN were on the same story in the same tone.

The next day the coin closed up 66%. The day after that it printed its all-time high. From the October 7 close to the October 9 intraday high is a 4.1x move in two days; on closing prices alone, 2.9x.

I can’t prove cause and effect. The timing is too tight to rule out coincidence, and something else may have been pushing it. But here is how I read it: for an asset like this, a mainstream “risk warning” is free national exposure. Half the people who saw the piece were annoyed by it. The other half looked up the name, found a coin attached to it, and bought.

That is why I keep watching it. None of the three runs came from inside crypto. The first was the videos. The second was the argument over whether he was a real person. The third was the press. Each one brought in people who had never heard of the token.

Five things I watch for a fourth run

One: whether the account is still producing, and whether the output has changed. This matters most and is hardest to check. Follower counts on both platforms aren’t visible in real time, so the only figures available are the ones printed on media pages: 267,700 on TikTok on October 8 (UNILAD), 244,000 on Instagram on October 14 (CreatorFlow), and 264,000 on Instagram plus just under 286,000 on TikTok on a BBC Bitesize page published in October, exact date unclear. The second set is higher than the first; if it was recorded a few days later, that means the account kept growing after the controversy, only more slowly. If he is still posting the same Paris-street shadowboxing, the idea isn’t used up yet. If brands, fashion weeks or real locations start appearing, that is a different thing: the account stops being a billboard for a token and becomes a character that can take work on its own.

Two: whether the first verifiable brand deal lands. One number here stings. The average rate for a sponsored post from a human influencer is $78,777. From an AI influencer it is $1,694. That gap is why brands are genuinely interested — the cost is more than forty times lower. The Spanish AI model Aitana López reportedly earns $75,000 to $100,000 a month. Jean Phil’s account posted something tagged “Paris Fashion Week” in early October and tagged Stella McCartney in it, but a tag is not a deal. I’m waiting for an announcement, not a tag.

Three: platform policy on AI labelling. Since August 31, 2026, Instagram has been reducing distribution for accounts that don’t label AI-generated people. For this character the policy cuts both ways. Add the label, and he admits he is fake — but he also gets a kind of legitimacy that lets a brand put money behind him. Leave it off, and he stays in the shadow of reduced reach. If the account switches on the “AI-generated profile” label itself, I won’t read that as surrender. I’ll read it as the signal that commercial work is starting.

Four: volume and new addresses. The main pool does about $243,000 a day against a $3.48 million cap. That is 7% turnover — only 7% of the market cap changed hands all day. Hardly anyone is trading. On the peak day it was $17.6 million of volume against an $11.2 million cap, a 157% turnover. The gap between 7% and 157% isn’t about price; it’s about how many people are involved. And where the new addresses come from is still unclear — genuinely new buyers, or old wallets passing tokens back and forth. That only becomes visible once it starts moving.

Five: whether the issuer touches that 1.4%. On-chain records show the creator wallet has never sold; the 13.6 million tokens bought for $44 are still there. The top ten addresses hold 16.5% between them, the largest single one 2.6%, and Rug.Tools shows liquidity fully locked or burned. Judged on contract permissions, this is not a project set up to run. Its risk isn’t in the contract; it’s that people stop talking about the character. Which makes that 1.4% the most direct signal available: the day he starts selling is the day he stops believing his own story.

One more thing, not a signal but a change of rhythm: copycats. A cluster of accounts has grown up around the character. “His wife,” Jeanne Misanthrope, reached 21,000 followers in days; another account under a near-identical name, Jeanne Philanthrope, has 147,600; there is Juanito Bravo; and there is a British character packaged as his boxing rival, Archibald Brown, who has a token of his own. The public record doesn’t say whether these are run by the same people as Jean Phil. If they start crossing over — booking a “fight,” appearing in each other’s videos — the cluster gets a second wind together, and only the original has a token attached, so money coming in would go there. If each stays in its own lane, it’s the same audience sliced a dozen ways, and nobody has enough.

Three ways this could go

The good one. The character lands a real brand deal, or gets a proper interview in traditional media as a case study in the AI-influencer economy, and the account goes back to posting often. Money that arrives then won’t be pricing the token alone; it will be pricing the character. On its own record, the third run reached $16.06 million, so a fourth run back into the $10–16 million range has precedent — but volume and address count have to move first, not price.

The likeliest one. The market cap keeps ranging between $2 million and $5 million. Every so often a video or an article pops it for two or three days before it falls back, with daily volume in the low hundreds of thousands. Nothing new happens. Fewer people pay attention each month.

The bad one. The account goes quiet, or the platform turns AI labelling into a hard reach penalty, or the 1.4% starts selling. This isn’t hypothetical — on September 21 it went from $11.2 million to $1.31 million in two days, and on October 2 the cap was back at $1.52 million. Getting from here to under $1 million takes a 71% fall, but it has done that before. For comparison: Dill Bits, the fake currency in Fortnite’s storyline, peaked at a $4.78 million cap and then fell 96%.

All three are scenarios, not forecasts. I can’t give you probabilities. Anyone who hands you “a 70% chance it runs” is making the number up.

What kind of asset this is, and how I’d hold it

There is no product and no revenue. The price depends on one variable: how many people are still talking about the character. And that pool of attention has already been shown to shrink. It ran three times, each rally higher than the last, and each time it gave the gains back within a few weeks. The reason is simple. Videos can be made indefinitely at almost no cost, but the number of people willing to watch is finite, and the same person watching a second time wants it less. The price follows that curve.

The contract itself has no obvious traps. Total supply is 969 million, all of it circulating — nothing locked, nothing to unlock later. The creator kept 1.4% and has never sold. The top ten addresses hold 16.5%. There is no website and no team on the record; the only things that count as official are a link in the Instagram bio and an X account called @JeanPhilMadame. So it is unlikely to go to zero overnight for a technical reason. It is far more likely to fade because nobody is discussing it anymore.

My approach is simple. If I buy, I buy a small amount as a ticket, not as a position. I don’t add while it is rising; I would only consider adding once it has proved that new money is coming in. From here, a fall of about 60% reaches the September 21 low, and a return to the peak needs more than 4.5x. Both directions are open, which makes position size matter more than the call.

Last word

I have been in this industry long enough to have watched a lot of “this IP is launching a token” stories end with both the IP and the token disappearing. What is different here is that the IP is real. Fifty million views are real. 244,000 followers are real. Even “people found out he was fake and wanted to talk about him more” is real. The only fake thing is the man.

As for the token, three rallies have taught me one thing: as long as people are still arguing about it, the price can be untethered from anything concrete. Three declines have taught me another: every leg up needs a new and larger piece of news to start it.

As for me — I’m Louise Lee, and I do research at Tronsell.io. Tronsell.io is an energy rental platform on TRON. We stake about 400 million TRX in our own pool and put out roughly 3.7 billion energy and 35 million bandwidth, serving more than ten institutional clients. What we normally do is the arithmetic that can be pinned down: on-chain fees, energy, bandwidth. What a person who doesn’t exist is worth, I can’t pin down. What I can do is check what can be checked, and admit there is no formula.

---

Prediction fact card (data as of November 2, 2026, 14:30, UTC+8)

Item	Detail
Token	Jean Phil (JEANPHIL)
Chain	Solana (SPL)
Contract	GTBxUiw6wJdmmkCGZgRHLyYxqu1vG4KtRpeox6yDpump
Main pool created	September 20, 2026, 03:06 (UTC+8), PumpSwap
Launch	Pump.fun; bonding curve filled in 51 minutes (pool migrated 21:06 Paris time, September 19)
On-chain supply	969,451,800, fully circulating, no unlock schedule
Price	$0.00357–0.00359 (CoinGecko $0.003571 / GeckoTerminal $0.003576 / DexScreener $0.003589)
Market cap	$3.46M / $3.47M / $3.48M (same three sources, same order)
24-hour volume	$243K (main pool) / $292K (all base-side pools) / $1.117M (CoinGecko aggregate) — see the note below
Base-pool liquidity	about $383,000 across five SOL-quoted pools, four with real depth, largest $290,000; plus two negligible pools quoted in other tokens
Depth ratio	about 11% (liquidity ÷ market cap)
Turnover	about 7% (24-hour volume ÷ market cap); 157% on the peak day
All-time high	$0.01656 (intraday, October 9, 2026), market cap about $16.06 million
All-time low	$0.00135 (September 21, 2026), market cap about $1.31 million
Low of this decline	$0.00238 (October 24, 2026), market cap about $2.31 million
Issuer holding	1.4% (13.6 million tokens, cost about $44); no on-chain record of a sale
Top ten addresses	16.5%; largest single address 2.6% (per Rug.Tools)
Website / team	no website, no disclosed team; official entry points are the bio link and the X account @JeanPhilMadame; operator unidentified
Centralised exchanges	no spot or futures listing announcement found

Volume differs sharply between the three sources. DexScreener shows $243,000 on the main pool and GeckoTerminal $292,000 across all base-side pools, and those two agree with each other. CoinGecko reports $1.117 million, close to four times higher. CoinGecko doesn’t say which pools it counts, so all three figures are listed here — for turnover, use the first two. CoinGecko’s number gives 32%.

The high also comes in two versions: the main pool’s daily candle records $0.01656 on October 9, while CoinGecko records $0.01525 at 07:20 UTC the same day — a gap of 8.6%. This article uses the first because it maps to a specific daily candle. The low diverges more. CoinGecko’s $0.00245 on October 24 is only the lowest point within its own coverage, and the $0.00135 candle from September 21 isn’t in it. That is why this piece keeps “low of this decline” and “all-time low” separate.

Sources

1. Origin: first Instagram video September 17, 2026 (shadowboxing inside an apartment); TikTok from September 19; first street clip 1.5 million views in three days; park clip 5.6 million views and 560,000 likes; one video shows 15.4 million views on the account page (a BBC page later gives just under 16 million; the two figures are from different dates).
2. Audience data: 267,700 TikTok followers on October 8 (UNILAD); 244,000 Instagram followers and 12 posts as of October 14 (CreatorFlow); a BBC Bitesize page published in October records 264,000 on Instagram and just under 286,000 on TikTok (the page says only “published October”; exact date unclear); L’ADN counts about 50 million views across the first seven posts; The Tab counts 13.8 million on the first video.
3. Identity and press: Know Your Meme lists the character as confirmed AI-generated; BBC, October 7, 2026 (reported by several outlets), describes a nonexistent influencer used to promote high-risk crypto assets, and a BBC Bitesize page adds that the creator told BBC Verify a token was not the original plan; Dexerto linked the AI character to token promotion on September 25; UNILAD on October 8 and CreatorFlow on October 14 each give a follower-growth timeline.
4. Launch: at 19:54 Paris time on September 19 a Pump.fun account renamed itself JeanPhilMadame; at 20:15 the token was created and 13.6 million tokens were bought in the same transaction (about 0.4 SOL); at 21:06 the bonding curve filled and the token migrated to PumpSwap (DexScreener records the pool at 19:06 UTC). The dollar value of 0.4 SOL has two versions: about $44 at the day’s SOL price near $111, and $43 as reported by UNILAD — a 2% difference. This article uses $44.
5. Price history: main-pool daily candles, September 19 to November 2, 2026 (GeckoTerminal), show three highs — $0.01158 intraday September 20, $0.01144 intraday September 25, $0.01656 intraday October 9 — and two deep lows — $0.0013466 intraday September 21 (all-time low) and $0.0023803 intraday October 24 (low of this decline).
6. Holdings and contract: the creator wallet holds about 1.4%; SoulTerminal shows no sale, and UNILAD likewise says “available data indicates the original creator has not sold”; Rug.Tools shows the top ten at 16.5%, the largest single address at 2.6%, and liquidity fully locked or burned; the first 20 buyers took 17.2% between them, and nine bundled addresses have been identified, now largely emptied. These figures come from Rug.Tools and SoulTerminal and may differ from raw on-chain records.
7. Issuer income: UNILAD reports more than $100,000 earned through Pump.fun fee sharing; Crypto Media Zone gives about $245,000. Neither figure has a second source, so it is treated as a range.
8. Copycat ecosystem: Jeanne Misanthrope (about 21,000 followers), Jeanne Philanthrope (147,600 followers), Juanito Bravo, and Archibald Brown, packaged as his boxing rival (with a separate token). Whether they share an operator cannot be confirmed from public material.
9. Platform policy: on August 31, 2026, Instagram renamed the “AI creator” label to “AI-generated profile” and began limiting recommendations for accounts showing AI-generated people without the label (Reels and Explore).
10. AI-influencer benchmarks: the average sponsored post from a human influencer is $78,777; from an AI influencer, $1,694. The Spanish AI model Aitana López reportedly earns 75,000–100,000 a month. Both figures are cited by UNILAD.
11. Comparable outcome: Dill Bits, the fake currency in Fortnite’s official storyline (February 2025), peaked at a $4.78 million market cap and then fell 96%; Epic and Tim Sweeney have both denied any connection.
12. Data: price, market cap, volume and price history from DexScreener, GeckoTerminal and CoinGecko, 14:20–14:40 (UTC+8) on November 2, 2026; total supply from a Solana block explorer.

Risk disclosure

This is a personal research note. It is not investment advice, and it is not a recommendation of any token. Memecoin prices are extremely volatile, and every figure here corresponds to a specific moment; by the time you read this, most will have changed. Whether JEANPHIL, the Jean Phil character and the social accounts promoting it belong to the same operator cannot be confirmed from public material. The project has no website, no disclosed team and no roadmap, trades only on decentralised exchanges, has limited liquidity depth, and its 24-hour volume is down about 98% from the peak. Only commit money you can lose entirely, and verify the contract address yourself before trading — there is more than one token using this name.