Polygon (POL): how the chain works, why it turned to payments, and its risks

Polygon crypto in 2026 means two things that are drifting apart. One is Polygon PoS, an Ethereum-compatible proof-of-stake network that adds a block about every 1.5 seconds and whose token, POL, pays for gas and secures the chain through staking. The other is Polygon Labs, the company that writes most of its software. Since mid-2025 it has shut its own zero-knowledge rollup, cut staff twice and agreed to buy two companies, Coinme and Sequence, for more than $250 million to build a stablecoin payments business.

The chain still carries a lot of dollars. DefiLlama counted $2.88 billion of stablecoins on Polygon on 29 October 2026, the eighth-largest amount of any network. Deposits in its DeFi apps were $695.7 million that day, down from $1.31 billion a year earlier, and one app, the prediction market Polymarket, accounted for about half of them. Live market data for the token is on our POL coin page.

What Polygon crypto is today: a chain, a token and a payments company

The Polygon name covers several products. Polygon PoS, which Polygon’s own 2026 posts call Polygon Chain, is the network most people mean. It runs the same smart contracts as Ethereum, so wallets such as MetaMask and apps written for Ethereum work on it with little change.

Around it sit three other pieces. Agglayer is a shared bridge that lets separate chains move assets between each other. The Chain Development Kit (CDK) lets a company launch its own chain that plugs into Agglayer. The Open Money Stack, unveiled in January 2026, is Polygon Labs’ payments product: bank deposits, wallets, routing between chains and payouts sold through one integration. A fifth product, Polygon zkEVM, is gone. Its sequencer was switched off on 3 July 2026.

Two organisations share the brand. A Polygon spokesperson told The Block in July 2026 that Polygon Labs “remains legally and structurally distinct from Polygon Foundation”. The Foundation looks after the network, the treasury and the ecosystem, and it releases protocol upgrades. Labs is the developer company. Its chief executive, Marc Boiron, has described the change as a move from a “blockchain foundation” to a “blockchain-enabled payments company”.

How Polygon PoS works: Bor, Heimdall and checkpoints on Ethereum

Polygon’s architecture documentation describes a two-layer system. Bor, a client built on Go Ethereum, produces the blocks and executes transactions. Heimdall-v2, built on the Cosmos SDK and CometBFT, is where the validators work: it watches the staking contracts on Ethereum, checks Bor’s blocks and periodically publishes a Merkle root of recent blocks on Ethereum. That record is a checkpoint, and it is what gives a withdrawal from Polygon to Ethereum its finality.

That design makes “layer 2” a loose fit. Staking and checkpoints live on Ethereum, but Polygon’s own validators decide which transactions are valid, and Ethereum neither re-executes nor proves them, as it would for a rollup. L2BEAT, which tracks Ethereum scaling networks, files Polygon PoS under “Others” rather than among rollups and describes it as a proof-of-stake sidechain. Polygon is still widely grouped with Ethereum’s scaling networks, and Coindaro files it there too. Our guide to layer 1 and layer 2 networks explains where the line usually sits.

Block production was rebuilt on 8 October 2025 with an upgrade called Rio. Instead of rotating through many validators, the validator set now elects a small pool of block producers; one builds blocks for a span while backups wait to step in. Rio also brought stateless validation, where nodes check new blocks without storing the full chain state, and a rule that splits transaction fees between the producer and the other validators. Polygon said Rio set the stage for about 5,000 transactions per second “in the near term”, against roughly 1,000 that it said earlier 2025 upgrades had reached on mainnet. That is the company’s target, not an independent measurement.

Smaller hard forks followed. Giugliano, activated on 8 April 2026, lets block producers announce blocks earlier; The Block reported that the change had cut finality by about two seconds on the Amoy test network. Ithaca, which Polygon announced as live on 30 July 2026, rotates in a new producer automatically when one stops. Before it, Polygon said, a failing producer could halt the chain for about 15 minutes with no automatic recovery. Its announcement put the fix this way: “Now a healthy producer will be rotated in automatically, so transactions keep moving.” (Polygon, 30 July 2026).

The chain’s current pace can be read from the chain itself. When we checked on 29 October 2026, the last 10,000 blocks had arrived on average every 1.50 seconds, each block could hold up to 160 million gas and the base fee was below a thousandth of a gwei. Users pay a priority tip on top of that.

Security comes from staked POL. Heimdall listed 104 validators in the active set on 29 October 2026, none of them jailed. The largest held 11.75% of total voting power, the top five 41.0% and the top ten 65.2%. A Validator.Info report on the Polygon forum ranked the biggest validators by stake on 1 March 2026: Upbit Staking with 384.8 million POL, Coinbase with 337.2 million and Binance Node with 255.3 million, ahead of Figment and Luganodes. Exchanges ran the top three.

From Matic Network to Polygon and POL

The project started in 2017 as Matic Network, founded by Jaynti Kanani, Sandeep Nailwal, Anurag Arjun and Mihailo Bjelic. On 9 February 2021 it announced its new name, Polygon, along with a plan to support many Ethereum scaling chains rather than one network. The token kept the MATIC name for three more years.

Two acquisitions followed in 2021. In August Polygon bought Hermez Network, a zero-knowledge rollup team, with 250 million tokens from its treasury, worth about $250 million at the time. Its 26 staff joined, and The Block reported that they would focus on zkEVM technology, the work that became Polygon zkEVM. In late 2021 Polygon added Mir Protocol in a deal Inc42 valued at about $400 million, paid in 190 million MATIC and $100 million of USDC. Renamed Polygon Zero, the Mir team published Plonky2, a proof system built for recursion, in January 2022.

In February 2022 Polygon raised $450 million in a round led by Sequoia Capital India that included a private sale of tokens. SoftBank, Tiger Global, Galaxy Digital and Dragonfly were among the other buyers named by TechCrunch. Nailwal gave the aim at the time: “The goal of Polygon is to become the AWS of Web3.”

On 4 September 2024 MATIC was upgraded to POL at one to one, and POL became the gas and staking token of Polygon PoS. Balances on the Polygon chain switched over automatically; MATIC held on Ethereum has to be swapped through Polygon’s portal. According to the token documentation, the swapped MATIC is held in the migration contract rather than burned, and governance can switch on a reverse conversion.

Of the four founders, only Nailwal is still in charge. Kanani and Arjun left their active roles about two years before Bjelic, according to The Block, and Bjelic stepped down from Polygon Labs and the Foundation board on 23 May 2025, writing: “I can no longer contribute to Polygon to the best of my abilities.”

POL tokenomics: about 2% new supply a year and one large burn

POL started with 10 billion tokens, one for each MATIC. The original design added 2% a year: 1% for the community treasury and 1% for validator rewards. A later proposal, PIP-26, raised the validator share for two years and brought it back to 1% from mid-2025. Polygon’s POL documentation gives the result as an effective 2% annual emission beginning after June 2025. New tokens are minted by an upgradeable EmissionManager contract, so governance can change the rate, but never above a ceiling written into the POL contract.

The POL contract on Ethereum reported a total supply of 10,740,293,555 POL when we read it on 29 October 2026. By Coindaro’s calculation, 2% of that supply comes to roughly 215 million new POL a year. There is no maximum supply. Market data sites can show different totals depending on how they count burned and bridged tokens.

Burns pull the other way. In September 2026, Polygon said in a post dated 29 September, the community completed a burn of 100 million POL from base fees collected earlier. Before the burn, The Defiant reported that later base-fee burns would become callable by the community each quarter. By our calculation, 100 million POL is 1% of the initial supply, a little under half a year of new emission at the October 2026 supply.

Whether the 2% should stay is an open fight. In October 2025 a forum member writing as VentureFounder proposed either setting inflation to zero at the next upgrade or cutting it by 0.5 percentage points each quarter, and sending at least 20% of the treasury’s quarterly net inflows to buying back or burning POL. The proposal suggested an on-chain vote in the first quarter of 2026. Its forum status still reads “Draft (Under Discussion)”. A former Polygon employee who co-wrote the POL design, qedk, answered: “Any proposal that eliminates inflation completely is untenable.” We found no vote and no new improvement proposal on emission, and the documentation still gives 2% as of 29 October 2026.

POL is a network token, not a share. Polygon Labs is a private company, and nothing Polygon has published ties the revenue of its payments business to POL, apart from burns paid for by chain fees. Our glossary entry on tokenomics explains how to read emission and burn figures like these.

What happened to Polygon zkEVM, and what users can still recover

Polygon zkEVM was the rollup the Hermez team had set out to build: a network that proves its transactions to Ethereum with zero-knowledge proofs. It never left its “Mainnet Beta” stage. On 11 June 2025 Marc Boiron, now chief executive of Polygon Labs, wrote on the Polygon forum: “We’ve made the decision to deprecate Polygon zkEVM Mainnet Beta next year.”

His post gave the reasons. Support for Ethereum’s cheaper data format (EIP-4844) took longer than planned. “ZK counters”, meant as a safety feature, made some DeFi transactions impractical. A custom client was hard for developers to debug, and the product did not stand out from other rollups. The time saved went to Polygon PoS and Agglayer.

The sequencer stopped on 3 July 2026 and the network produces no more blocks. People who held assets in their own wallets on zkEVM can recover them on Ethereum through a Claims interface that Polygon says stays open until 31 December 2027. Assets that sat inside smart contracts at shutdown, such as liquidity positions, multisig wallets or third-party bridge balances, cannot be recovered through that tool; Polygon tells their owners to contact the protocol that holds them. DefiLlama still listed about $46.3 million in the Polygon zkEVM bridge contract on Ethereum on 29 October 2026. That figure does not say whose funds they are or how much of it can still be claimed.

Agglayer, CDK and Katana: the cross-chain plan that survived

Agglayer is what remains of the 2021 idea of many chains under one roof. Chains connected to it share one bridge. Since 3 February 2025 that bridge has used “pessimistic proofs”, which Polygon says ensure that no connected chain can withdraw more than has been deposited into it. The point is that a weakly secured chain should not be able to drain funds that belong to its neighbours.

On 19 October 2026 Polygon added Hyperledger Besu to its CDK. A bank or a group of institutions can now run a private chain with its own validators, keep transaction data in-house and still be connected to Agglayer from its first block. Polygon notes that the proofs cover bridge accounting, not every transaction inside the private chain. No bank had been named as a user when we checked.

Katana is the best-known chain born from this plan. Polygon Labs and the market maker GSR incubated it, and it opened its public mainnet at the end of June 2025 as a DeFi chain built on a custom version of the OP Stack and connected to Agglayer. Its KAT token has a fixed supply of 10 billion, with about 15% set aside for POL stakers. At launch Nailwal said Katana held more than $240 million of deposits. DefiLlama measured $36.1 million locked on Katana on 29 October 2026.

The payments turn: Open Money Stack, Coinme and Sequence

On 13 January 2026 Polygon Labs signed agreements to buy Coinme, a crypto exchange founded in 2014, and Sequence, a wallet infrastructure firm founded in 2017, for more than $250 million in total, The Block reported. The two were to play “foundational roles” in the Open Money Stack, unveiled the week before. The stack is meant to let a business take money in by bank transfer, card or cash, hold it as stablecoins, move it across chains and pay it out, all through one integration.

The Coinme deal closed on 14 October 2026. According to Polygon, its US payments business now covers 48 states with 37 state money transmitter licences and 11 exemptions, and is registered with FinCEN as a money services business. Four launch customers were named: Tazapay for cross-border business payments, Mesh for funding accounts from existing wallets and exchanges, Toku for payroll, and Kira for stablecoin and banking services to fintechs.

The stack is not tied to Polygon’s own chain. Polygon Labs added TRON in early October 2026 and Solana on 29 October, as our news report on the Solana launch details. For POL holders this is the central tension of the strategy: the company’s revenue can grow from payments that never touch Polygon PoS.

Polygon’s chain does have payment traffic to show. In a post dated 29 September 2026, Polygon said stablecoin transfers on the chain had passed $3 trillion in total since September 2020, with $933 billion in 2025 and $741 billion in 2026 up to that date. It named Revolut, Paxos, Polymarket, Cash App, Deel and Tazapay as users, said Revolut moved $810 million on Polygon in 2025 and said PayPal USD began issuing natively on the chain this summer. On 29 April 2026 Polygon announced that Meta now pays select Facebook creators in Colombia and the Philippines in USDC, over Polygon or Solana. Transfer volume counts the same dollar each time it moves, so it measures activity, not how much money stays on the chain.

The turn has cost jobs. Polygon Labs cut about 60 people in January 2026, The Block reported, and Boiron announced a second round on 16 July 2026 without giving a number. He tied the merger work to a goal of being profitable in 2027 and told staff: “Two rounds of changes in one year is a lot to ask of a team” (The Block, 16 July 2026). Earlier cuts took about 100 people in February 2023 and 60 in 2024.

Who runs Polygon: Labs, the Foundation and a 13-seat council

Power over the network was concentrated in June 2025. On 11 June, the day the zkEVM shutdown was announced, Nailwal became the first chief executive of the Polygon Foundation, with full executive control. Cointelegraph described the step as a shift away from the decentralised governance model the Foundation had embraced. His line that day: “Now, it is time to again move fast, aggressively and with full conviction and focus.” Boiron runs Polygon Labs; he posted the zkEVM closure notice and announced the July 2026 cuts.

Changes to Polygon’s system contracts go through the Protocol Council, a 13-signer multisig proposed in PIP-29 in October 2023. Under that proposal, regular changes need 7 of 13 signatures and a 10-day delay, and emergency upgrades need 10 of 13 with no delay. PIP-77, posted on 2 February 2026, lists the 13 proposed signers for 2026, among them L2BEAT, Mehdi Zerouali of Sigma Prime, Vahe Karapetyan of Hexens, Liz Steininger of Least Authority, Viktor Bunin of Coinbase and two multisig signers from Polygon Labs’ engineering and security teams. The forum page shows no ratification note. Sigma Prime, Hexens and Least Authority all appear on Polygon’s list of auditors, so firms that review the code also sit among the people who approve changes to it.

Security record: audits, two bugs in 2021 and a finality delay in 2025

Polygon’s security reports page lists audits by Informal Systems for Bor and Heimdall, by ChainSecurity and Sigma Prime for the POL token, and by Sigma Prime, Hexens and Spearbit for the Agglayer contracts; most entries carry no date. ChainSecurity’s report on the POL token, completed on 4 October 2023, concluded that the code “provides a high level of security”. The same page says Polygon Labs has held ISO/IEC 27001 certification for its information security management since March 2024.

The worst moments came in 2021. On 5 October that year a whitehat, Gerhard Wagner, reported a flaw in the Plasma bridge that would have let a single withdrawal be replayed up to 223 times, with about $850 million at risk. Polygon paid him $2 million and no loss was reported, according to Immunefi. In December a second bug, in the genesis contract of Polygon PoS, put more than 9.2 billion MATIC at risk. An attacker took 801,601 MATIC before an emergency hard fork on 5 December closed the hole; Polygon said it would cover the loss and paid about $3.46 million in bounties, The Block reported.

On 10 September 2025 the chain kept producing blocks but stopped finalising them on some nodes. Some nodes stalled, apps using affected RPC providers were disrupted, and a few providers had to rewind to block 76,273,070, according to the Polygon status page. The lasting fix was a Heimdall hard fork scheduled for about 15:00 UTC the same day, and the incident was closed the next morning. We found no successful attack on Polygon PoS consensus or on the POL contracts since December 2021. Apps deployed on the chain are a separate matter, with their own record.

Polygon by the numbers: TVL, stablecoins and fees on 29 October 2026

All figures below come from DefiLlama on 29 October 2026, except the validator count (Heimdall) and the block time (our read of the chain the same day).

From those DefiLlama figures, Coindaro calculates that Polymarket International produced about 84% of the fees paid on Polygon in the 30 days to 29 October 2026 ($32.35 million of $38.6 million) and held about 52% of the chain’s TVL ($360.8 million of $695.7 million). The chain’s activity therefore leans on one app; our Polymarket project page covers that app and its own risks. The network’s own fee line, $2.11 million in a month, is small next to what apps charge their users.

Stablecoin counts also vary by method. A Polygon spokesperson gave The Block a figure of $3.37 billion in July 2026, against DefiLlama’s $2.88 billion in late October. Polygon ranked eighth among chains in both.

Polygon crypto vs Solana, Base and Arbitrum: smaller on deposits and stablecoins

On the same DefiLlama data, Solana had $5.69 billion locked and $15.25 billion of stablecoins, Base $6.11 billion and $5.07 billion, and Arbitrum $1.51 billion and $3.47 billion. TRON, the main rail for USDT transfers, held $93.6 billion of stablecoins. Polygon trails all four on both measures.

Its case rests elsewhere: very low fees, compatibility with Ethereum tools, a US-licensed money business through Coinme and a list of payment users named by Polygon. The company has hedged its bet, though. By plugging Solana and TRON into the Open Money Stack, Polygon Labs sells payments on its rivals’ networks too.

Risks to weigh, and why we rate Polygon 3 out of 5

We give Polygon a risk score of 3 on our 1 to 5 scale. The chain has carried stablecoin transfers since 2020, has come back from each incident in its own code listed above and serves named payment companies. Against that, several risks are specific to it.

The company carrying most of the development has cut staff twice in a year, shut a flagship product and staked its future on payments, where it competes with banks and processors. POL holders have no claim on that business.

Activity is concentrated. By our calculation, one app produced about 84% of fees in the 30 days to 29 October, and any problem at Polymarket would show up in Polygon’s numbers at once.

Supply grows by about 2% a year, and the proposal to end that has had no vote in a year. Burns depend on the fee income the chain collects, which is small today.

Control is narrow. Three exchanges were the largest validators in March 2026, ten validators held 65.2% of voting power in October, the Foundation chief has full executive control, and contract upgrades pass through a council that includes Polygon Labs signers.

Frequent hard forks, such as Giugliano and Ithaca in 2026, keep improving the chain, but each one is a chance for a bug, as the September 2025 finality delay showed. Former zkEVM users with funds held in smart contracts cannot use the Claims tool. Stablecoin payments also depend on issuers and on rules still being written in the US and Europe.

Next dates: quarterly burns, PIP-77 and the zkEVM deadline

We found no hard-fork date announced for the rest of 2026. Polygon’s latest Bor client releases already contain pre-confirmations, a way for users to get a fast promise that a transaction will be included, but the feature ships switched off by default since Bor v2.10.2 of 30 September. The next base-fee burn can be triggered by the community each quarter, the inflation proposal still has no vote, and PIP-77 has no ratification note on the forum. Polygon Labs aims to be profitable in 2027, and the zkEVM Claims interface closes on 31 December 2027.