← All articles

DPoS and the 27: How TRON's Super Representatives Work

Key takeaways

  • TRON's blocks all come from 27 elected addresses — Super Representatives — that rotate through 3-second slots. You can't run a node and join; you have to win an election.
  • Candidates rank 1–127 by vote weight: the top 27 produce blocks, the next 100 (SR Partners) earn voting rewards only, and the boundary moves every six hours.
  • SRs pay rewards out as periodic mass-outbound TRX transfers — a normal pattern that naive tooling mistakes for fund dispersion.
  • Vote concentration means governance concentration: 18 of 27 SRs pass a proposal, so 10 can veto one.

The history article covers Delegated Proof of Stake in a paragraph — right for an overview. This goes deeper: how DPoS actually runs, who holds the 27 block-producer seats, how the election works, and why all of it shows up when you read TRON transaction data.

The short version: TRON concentrates block production in a small, elected group, and that concentration is visible on-chain. Knowing who’s in the group, how they got there, and what their addresses look like in the transaction graph separates infrastructure noise from real user behavior.

Three ways to agree on a block

Public blockchains pick one of three consensus families, each trading throughput against decentralization.

DPoS was formalized by Dan Larimer — it’s also behind BitShares and EOS — and it’s the consensus under TRON. The trade-offs fall out of the committee size. A small, known group coordinates fast, so TRON lands a block every three seconds, which PoW can’t touch. But 27 producers is far more concentrated than Ethereum’s validator set; supporters call that fair because the committee is elected, while critics point out that the same entities tend to hold seats indefinitely as votes accumulate. The upside on governance: because producers are known addresses with declared identities, protocol changes get voted on explicitly instead of signaled through miner behavior.

For TRON, the practical consequence is that block production isn’t permissionless. Running a node earns you nothing. You have to win the vote.

The 27 active Super Representatives

At any moment, exactly 27 addresses hold the Super Representative (SR) title and produce every block on TRON mainnet. The number is fixed in the protocol.

Each three-second window is a slot, and the 27 active SRs cycle through them on a rotating schedule. If the SR assigned to a slot doesn’t produce — downtime, network trouble, or deliberate skipping — that slot is missed and the next SR takes the next one. The chain doesn’t stall.

+3s +3s +3s +3s SR #1 SLOT SR #2 SLOT SR #3 MISSED SR #4 PRODUCES SKIP · CHAIN CONTINUES 28,800 slots/day (86,400s ÷ 3) → ~1,067 blocks per SR per day
Slots rotate every three seconds; a miss skips forward, and the chain continues.

With 28,800 slots in a day (86,400 seconds ÷ 3), an even split puts each SR around 1,067 blocks daily — exact counts drift a little with missed slots and rounding.

Block production pays out two ways:

Both per-block figures are network parameters set by governance — an SR proposal can change either — so read them as current documented values, not fixed constants.

Each SR sets a brokerage rate (or commission rate) that splits rewards between its own address and its voters. A newly elected SR defaults to 20% — keep 20%, distribute 80% pro rata to voters — and can change it later. SRs courting voter support tend to run lower; SRs backed by entities that vote for themselves, like large exchanges, often sit at the default or higher.

That distribution is a batch on-chain operation: it lands as mass outbound transfers from the SR address. To tooling that doesn’t know TRON’s reward mechanics, an address firing dozens or hundreds of small TRX transfers each epoch can read as suspicious dispersion. It’s almost always an SR paying its voters.

SR Partners — the bench behind the 27

SR status isn’t binary. TRON keeps a ranked candidate list, and the eligible set runs from rank 1 to 127.

Ranks 1–27 · Active Super Representatives Produce every block. Earn block rewards (8 TRX/block) + voting rewards. Self-vote and seats both shift each election. Ranks 28–127 · SR Partners No blocks. Voting rewards only, at their brokerage rate. The promotion bench. Ranks 128+ · SR Candidates Registered (WitnessCreateContract, 9,999 TRX). Vote-eligible, earning nothing. RANK 127 VOTES
Vote weight sorts candidates into three tiers; only the top 27 produce blocks.

Partners are the promotion bench. As vote weight shifts, Partners get promoted into the active 27 and sitting SRs get displaced — the rank 27/28 boundary can move every election. Partner addresses sit in the same reward-distribution chain too, so an address running mass outbound TRX distributions every six hours could just as easily be a Partner as an active SR.

Getting onto the candidate list costs a registration fee of 9,999 TRX, paid on-chain as a WitnessCreateContract transaction. Finding that transaction in a wallet’s history is a real identity marker: the address registered as TRON infrastructure, or at least paid to look like it.

Voting runs on TRON Power

Elections never stop — they settle on a clock. Every six hours TRON runs a maintenance period: it tallies the votes, updates the SR rankings, and hands block production to the new top 27. Informally this window gets called an “epoch.”

Tally TP votes EVERY 6H Re-rank 1–127 SET TOP 27 New SRs produce BLOCKS Pay rewards MASS OUTBOUND NEXT MAINTENANCE PERIOD
The six-hour cycle: tally, re-rank, produce, distribute, repeat.

Votes are denominated in TRON Power (TP). Under Stake 2.0, one TRX staked for votes yields one TP, which grants one vote. TP is non-transferable and non-tradeable — it lives on the staking address, and unstaking the TRX removes the TP and its votes at the same time.

The voting act itself has no lock-in. A voter can reallocate votes any time between maintenance periods, and the change takes effect at the next election. Only the staked TRX faces the 14-day unstaking wait introduced by Stake 2.0 (TIP-467, activated April 2023); the voting direction moves freely. So an SR’s vote count can swing hard inside one six-hour window if a large holder shifts allocation.

Voters can also split TP across several candidates. Exchanges and staking services routinely spread votes to hold multiple seats at once, or to prop up Partners sitting near the rank 27 boundary.

Running a node earns you nothing on TRON. You have to win the vote.

Who actually holds the seats

Three kinds of entity dominate the active set.

Major exchanges. Exchanges custody large amounts of user-deposited TRX. Run an SR node, vote that custodied TRX for yourself, and you collect block and voting rewards on assets held for customers — economically rational, and explicitly permitted by the protocol. Binance Staking has historically ranked among the top vote-getters. The Block’s September 2019 report found Binance’s candidacy captured over 59% of all network votes — and that a single Binance cold wallet, one of roughly 650 addresses voting for it, supplied nearly 12 billion TRX, over 99.60% of the candidacy’s own vote total. Same report, two denominators: a majority of every vote cast on the network, and near-total concentration within Binance’s own supporters. OKX, HTX (formerly Huobi), and Poloniex have also held active seats. Exact rankings move with market conditions and custody volumes.

Dedicated staking services. Operators like P2P.org (elected as an SR in April 2025) exist to run professional node infrastructure. They typically offer voters a revenue share, and sometimes run lower brokerage rates to attract delegation.

Foundation- or Justin-Sun-affiliated entities. TRON’s early SR set included addresses tied to the Foundation. That share has fallen as the network grew, but anyone examining network control should read the current roster and address annotations rather than trust a static description.

Because large holders self-vote and exchanges hold the biggest TRX concentrations, the set leans toward institutional incumbency. Community nodes show up in the Partner tier, but holding an active seat takes vote weight that community operators rarely sustain against institutional competition. This isn’t a TRON quirk — concentration is a structural feature of every DPoS system. For the live roster with current vote counts, TronScan’s SR page is the authoritative source.

Governance: two tracks, one veto

Changing the TRON protocol runs through a spec track and an on-chain track.

TIPs (TRON Improvement Proposals) are the community spec layer. Anyone can author one, starting as a GitHub issue in tronprotocol/tips and moving to a formal pull request. TIPs progress through Draft, Last Call (at least two weeks of public review), and Accepted. An accepted TIP that needs a protocol change then goes on-chain as a network proposal.

Network proposals are the activation layer. Only SRs, SR Partners, and SR Candidates can submit one. If at least 18 of the 27 active SRs vote yes, it passes and takes effect automatically. That 18-of-27 bar — two-thirds plus one — means 10 SRs can block any change. Vote concentration and veto power are the same thing: whoever dominates the set by vote weight holds the off-switch.

Three TIPs show the pipeline in action:

All of this is visible on-chain. Proposal creation and SR vote transactions cluster on SR addresses around upgrades. They’re ordinary artifacts of SR status rather than evidence of coordination.

Why it matters when you read the chain

The SR system throws off predictable patterns that anyone working TRON data will hit. Recognizing them kills false positives.

WitnessCreateContract transactions register an SR candidate. One in a wallet means the address self-registered — either it belongs to an entity running TRON infrastructure, or someone paid 9,999 TRX to look like one. Cross-reference TronScan’s SR history to see whether it was ever elected.

Periodic mass-outbound TRX transfers are voting rewards. SRs and Partners pay delegators every maintenance period, producing a rhythmic burst of dozens to hundreds of small TRX transfers, sized to each recipient’s vote share. A system that flags “mass outbound transfers” as suspicious will trip on this. The tells that clear it: six-hour periodicity, proportional amounts, and a sender that’s a registered SR or Partner.

Resource delegation from SR addresses is common — SRs delegate Energy or Bandwidth to their own operational wallets or partner addresses. An SR at the origin of a delegation chain doesn’t imply it controls the recipient; it usually reflects a service relationship, like subsidizing energy for a staking product’s users.

Vote and block-producer traffic. VoteWitnessContract transactions come from voting addresses; a high volume from one address points to an exchange, staking service, or large holder managing allocations — not manipulation. And the block-producing SR is recorded in every block header, so if you need to know which SR was active in a given window, the header answers directly.

This is why SR-affiliated addresses should be read as infrastructure when tracing a wallet’s funding history rather than treated as genuine sources of funds: their outbound transfers are protocol reward distributions, not deliberate funding relationships. Any funding-origin analysis on TRON has to apply that filter, or it will read routine SR reward payouts as real financial backing.

Sources

Primary documentation and data sources used for facts in this article: