How TON Is Distributed Across the Ecosystem Today
We collected every TON address holding 100 TON or more into our database and refreshed all their balances in a single pass. That gave us 100 982 wallets holding 5 278 655 753,78 TON. Here is how that amount is distributed—by wallet size, and by who owns it.
In brief
- 100 982 wallets hold 5 278 655 753,78 TON
- Three wallets account for half of the entire amount
- 80,47 % of wallets fall into the smallest tier (100–1 000 TON) and hold 0,45 % of the TON
- 49 wallets holding 10 million TON or more control 88,01 %
- 24,23 % of the total sits on addresses with
uninitstatus—they have never been initialised - The Gini coefficient for this snapshot is 0,9936
What the snapshot covers
The 100 TON floor is deliberate. Below it, addresses turn into noise: airdrop dust, empty application wallets, leftovers after a withdrawal. The 100 982 addresses in the snapshot cover practically the entire network: against a 5,5 billion TON total-supply reference, these addresses account for roughly 96 %. Everything smaller than 100 TON is scattered plankton that does not move the distribution picture.
Balances were synchronised on 23 October 2026—all 100 982 addresses refreshed in one batch.
Tiers: from shrimp to leviathan
We split the wallets into seven steps, each one ten times larger than the last.
| Tier | Balance | Wallets | % of wallets | TON | % of TON |
|---|---|---|---|---|---|
| 🦐 Shrimp | 100 – 1 000 | 81 256 | 80,47 % | 23 929 394 | 0,45 % |
| 🐟 Fish | 1 000 – 10 000 | 16 249 | 16,09 % | 45 385 193 | 0,86 % |
| 🐬 Dolphin | 10 000 – 100 000 | 2 781 | 2,75 % | 74 795 751 | 1,42 % |
| 🦈 Shark | 100 000 – 1 M | 517 | 0,51 % | 146 118 753 | 2,77 % |
| 🐋 Whale | 1 M – 10 M | 130 | 0,13 % | 342 837 135 | 6,49 % |
| 🐳 Humpback | 10 M – 100 M | 44 | 0,04 % | 1 554 013 844 | 29,44 % |
| 🌊 Leviathan | 100 M and above | 5 | 0,005 % | 3 091 575 684 | 58,57 % |
Two charts of the same data, mirrored against each other. First, how many wallets sit in each tier:
And now how much TON each tier holds:
The four bottom tiers are 99,82 % of all wallets and 5,50 % of the money. The median wallet in the snapshot holds 272,66 TON; the average holds 52 273,23 TON. That 192× gap between median and mean is the whole story of this distribution.
Three wallets are half the network
The concentration here is harsher than any familiar comparison.
Seen from the other side:
The remaining 90 982 wallets—all of them together—hold 0,73 % of the total.
The Gini coefficient across the snapshot is 0,9936, against a theoretical maximum of 1,0. For comparison, the most income-unequal country in the world does not reach 0,65. But a blockchain is not supposed to behave like a national economy: half of the top here is not anyone’s personal money, it is protocol contracts. Which addresses exactly—below.
The network’s top 15 wallets:
| # | Wallet | Category | Balance, TON | Share |
|---|---|---|---|---|
| 1 | Elector Contract | system-contract | 1 382 625 743 | 26,19 % |
| 2 | TON Believers Fund | treasury | 1 245 901 482 | 23,60 % |
| 3 | Ton Ecosystem Reserve | reserve | 248 223 656 | 4,70 % |
| 4 | Miner Cluster #MBBPEE | miner | 112 400 335 | 2,13 % |
| 5 | Telegram | organisation | 102 424 468 | 1,94 % |
| 6 | Miner Cluster #Hm73vA | miner | 99 337 701 | 1,88 % |
| 7 | Fragment · Telegram Ads | organisation | 94 281 115 | 1,79 % |
| 8 | Miner Cluster #mQ2-8p | miner | 87 810 798 | 1,66 % |
| 9 | Miner Cluster #q4praA | miner | 69 275 202 | 1,31 % |
| 10 | Binance | exchange | 59 570 010 | 1,13 % |
| 11 | Miner Cluster #lfvnRz | miner | 58 445 004 | 1,11 % |
| 12 | Miner Cluster #79ll7x | miner | 58 243 498 | 1,10 % |
| 13 | Miner Cluster #bsoXrR | miner | 54 271 398 | 1,03 % |
| 14 | Miner Cluster #DSfAHB | miner | 46 215 000 | 0,88 % |
| 15 | Miner Cluster #YNVvHn | miner | 42 722 047 | 0,81 % |
Ten of the fifteen largest addresses are early-mining wallets.
Who holds TON: by category
Every wallet in our database can be assigned a category from a controlled list: system-contract, treasury, reserve, organisation, miner, exchange, validator, bridge, staking, marketplace, burn, personal. A category is set when the wallet has confirmed labelling; the rest simply have none.
In absolute numbers:
| Category | Wallets | TON | % of TON |
|---|---|---|---|
| miner | 58 | 1 494 946 137 | 28,32 % |
| system-contract | 5 | 1 391 278 469 | 26,36 % |
| treasury | 2 | 1 245 908 994 | 23,60 % |
| no label | 100 262 | 461 131 892 | 8,74 % |
| organisation | 40 | 263 783 702 | 5,00 % |
| reserve | 2 | 248 224 376 | 4,70 % |
| exchange | 195 | 142 957 458 | 2,71 % |
| unclassified | 375 | 15 960 821 | 0,30 % |
| bridge | 12 | 11 131 051 | 0,21 % |
| staking | 10 | 1 166 096 | 0,02 % |
| validator | 12 | 912 549 | 0,02 % |
| marketplace | 6 | 777 842 | 0,01 % |
| personal | 2 | 465 654 | 0,01 % |
| burn | 1 | 10 713 | 0,0002 % |
The top three rows are 78,28 % of the entire amount, spread across 65 addresses.
System contracts hold 1,39 billion TON, and almost all of it is Elector: 1,38 billion staked by validators to take part in consensus. These coins belong to no one personally—they are the network’s working collateral. The same category holds .ton DNS with 7,96 million, the Config Contract, and the zero address with 10 713 burned TON.
Treasury and reserve account for another 1,49 billion across three addresses: TON Believers Fund (1,25 billion) and Ton Ecosystem Reserve (248 million).
Organisations are 40 wallets holding 264 million. That includes Telegram (102 million), the Fragment and Telegram Ads wallets (134 million across four addresses), Wallet in Telegram, and TON Foundation (OLD).
Exchanges are 195 addresses holding 143 million TON, or 2,71 %. The largest is Binance with 59,6 million, followed by OKX and four Bybit wallets at roughly 15 million each.
For a network that trades on centralised venues, that is not much: exchanges hold almost nine times less TON than the frozen early-mining layer.
1,22 billion TON in wallets that were never opened
TON’s most striking anomaly is early mining. 58 addresses carrying the miner category hold 1,49 billion TON—more than a quarter of the entire amount. But the state of those wallets is more interesting than the volume.
43 of the 58 addresses carry uninit status: coins sit on them, but the wallet code has never been deployed. In other words, the private key for these addresses has not been used once since the reward landed. 1 224 380 977 TON—23,2 % of the entire snapshot—sits in wallets nobody has ever opened.
Counting across all categories, there are 463 such addresses holding 1 279 057 644 TON (24,23 %). Nearly a quarter of the network technically exists but does not move.
One caveat: status was not checked for every address—89 492 of them have it empty. But together those addresses hold only 150 million TON (2,84 %), so they do not affect the picture for the large balances.
The blank spot: 99,3 % of wallets carry no label
Now the uncomfortable part. Only 720 wallets out of 100 982 have a confirmed category—that is 0,71 %. The other 100 262 addresses in the snapshot are unlabelled.
That sounds like a failure until you look at the money: those 720 wallets hold 91,26 % of the total. And if you drop the 375 addresses that were reviewed and marked unclassified, 345 wallets remain with a substantive category—holding 90,96 % of the TON.
Labelling coverage rises in lockstep with balance:
All 49 wallets holding 10 million TON or more are fully reviewed. In the whale tier (1–10 million) one in four is labelled, and this is where the most interesting blank spot sits: 97 unlabelled wallets holding 202,5 million TON. The largest of them are early mining too—clusters of 4–5 million TON that simply have not been taken through to a confirmed category yet:
| Wallet | Status | Balance, TON |
|---|---|---|
| Early Miner Clusters #Mwdpu3 | active | 5 204 097 |
| Early Miner Clusters #Bmkn5_ | active | 5 100 201 |
| Early Miner Clusters #mWzTbd | active | 5 100 000 |
| Early Miner Clusters #1pNV3f | active | 5 000 001 |
| kraken | active | 4 408 800 |
In the bottom tier only 0,24 % of wallets are labelled—and they do not need to be: all 81 256 shrimp together hold 0,45 % of the TON. Labelling works top-down by balance, and in money terms it is essentially done.
The 32 unlabelled wallets inside the top 100 stand out separately—they hold 114 million TON. That is the next target.
What this adds up to
TON’s distribution is not one curve but three distinct layers that together produce that alarming Gini coefficient:
- The protocol — 2,88 billion TON (54,7 %) in Elector, the treasury and the reserves. These coins are not on the market: part is staked by validators, part sits in foundation contracts.
- The early-mining freeze — 1,22 billion TON (23,2 %) in wallets that were never initialised. Technically that is another non-circulating slice.
- Everything else — roughly 1,17 billion TON: exchanges, organisations, bridges, staking, and a hundred thousand ordinary wallets.
In other words, the genuinely liquid part of the network is about a fifth of what the headline supply suggests. And 80 % of the addresses in our snapshot are wallets holding 100–1 000 TON, which together do not add up to a fortieth of Elector’s balance.
Sources and methodology
All figures are aggregates from our own database, with no external sources.
The snapshot. The wallet table, rows with balance >= 100 TON: 100 982 addresses. The list was built by walking TonCenter’s topAccountsByBalance in descending balance order down to the 100 TON floor, and balances were rewritten in one batch on 23 October 2026. Balances are stored in nanotons and converted to TON by dividing by 10⁹.
Tiers are our own decimal-step split. Boundaries include the lower value and exclude the upper one—a wallet holding exactly 1 000 TON lands in “fish”.
Categories come from the wallet_classification table, rows with kind = entity and status = approved. When a wallet has several, the highest-confidence row wins—the same logic that drives the category shown on a wallet’s page. Rows originate from the ton-labels import, our classification model, and manual moderation. “No label” in the table means there is no approved row at all; unclassified means the wallet was reviewed and no substantive category was found.
The active / uninit statuses come from the status field of the wallet table, holding the TonAPI value as of the last account check. uninit means the address has a balance but no deployed wallet code. Status was not checked for every address in the snapshot: 89 492 have it empty, and they are shown as a separate row in the status breakdown.
The Gini coefficient is computed over the balances of the 100 982 wallets in the snapshot. Addresses below 100 TON are excluded—including them would push the coefficient higher still.
Data boundaries. This is a distribution across addresses, not across people or organisations. One owner can hold dozens of addresses, and a single address—an exchange wallet or a staking contract—can hold the funds of thousands of users. The snapshot reflects the state of our database as of 23 October 2026, not the entire blockchain.