Threshold Network (T): the tBTC bitcoin bridge run by 100 signers per wallet, and a pause with no confirmed cause

Threshold Network tBTC is a token on Ethereum meant to be worth one bitcoin, with the coins held not by a custodian but in wallets that 51 of 100 node operators must agree to sign for. As of 1 November 2026, nobody can start a mint or a redemption through the Threshold app: new deposits, minting and new redemption requests stopped around 30-31 October, Aave has frozen its tBTC reserve on Ethereum Core, and the cause of the pause has not been explained in anything we could read.

T is the separate governance token of the network, which came out of the merger of Keep and NuCypher. It has 11.155 billion units, no further emissions, and gives holders votes and a fee discount rather than a share of income. tBTC itself held $336.5 million on DefiLlama at about 19:28 UTC on 1 November 2026, against $661.0 million a year earlier. This page explains how the bridge works, who can change it and what is still unknown. For the day-by-day account of the pause, read our news report on the tBTC pause and Aave’s freeze.

What Threshold Network tBTC is

A wrapped bitcoin lets BTC be used in Ethereum apps such as lending markets. The usual design is simple: a company holds the coins and issues a token. WBTC and Coinbase’s cbBTC work that way. tBTC replaces the company with a group of software-run signers, and our guide to how cross-chain bridges work puts that choice in context. Threshold’s docs say the design means “no single entity can act alone on funds”. That sentence is true of one signer. It says less about the group of operators and the committee above them, as the sections below show.

The project has two parents. Keep, developed with the studio Thesis, built the first tBTC, released in 2020. NuCypher worked on threshold encryption. Their communities voted to merge in 2021, and Threshold says the final proposal, RC0, “passed with an overwhelming majority vote”. T followed around the start of 2022. The funding we could trace dates from before the merger: $10.67 million for NuCypher in October 2019 and $7.7 million for Keep in April 2020. tBTC v2, the version in use today, was announced as live on 31 January 2023, with a permissioned group of minters and guardians in charge of the early phase.

Since February 2025 the day-to-day work has moved to a company. DAO votes TIP-098 and TIP-100 created Threshold Labs (tLabs) to lead product, development and marketing for tBTC. Around the bridge sit the Unified Bitcoin Router, a single app for minting, redeeming and bridging, and Verifiable Bitcoin Accounts, an institutional product announced on 23 April 2026, before deployment, that is meant to keep each account’s coins apart instead of pooling them. The older thUSD stablecoin is in maintenance mode.

How tBTC works: 100 signers, 51 needed

About every 14 days the system creates a new Bitcoin wallet, the interval the Bridge contract still reports. The docs on wallet generation describe the steps: a Sortition Pool picks 100 signers, using randomness from a Random Beacon, and any spending needs 51 of them to take part in a threshold ECDSA signature. No full private key exists in one place.

A hundred seats does not mean a hundred independent parties. Each seat is drawn on its own, so one operator can hold several. On 1 November 2026 we read the pool contract on Ethereum, which listed 20 operators. That is a registration count: it does not show how many nodes are online or who runs them. In November 2024 a DAO proposal counted 35 approved addresses run by 21 distinct entities. The docs’ own example shows why concentration matters: a staker with 35% of the weight would have about a 0.074% chance of controlling one wallet, and about 3.8% over 52 wallets.

The list is also closed. Threshold’s operator page says “Signer nodes form a permissioned set”. Operators need Linux machines kept up to date with security patches and uptime above 96%, and should be able to upgrade within 24 hours of notice.

Minting and redeeming tBTC

A depositor sends BTC to a Bitcoin address that carries a time-locked refund, nine months in the docs. Roughly every eight hours the signers “sweep” recent deposits into the wallet’s main output, and a Bitcoin SPV proof shows the contracts on Ethereum that it happened. Then tBTC is minted. Redemption runs the other way: the holder gives up tBTC and names a Bitcoin address, and the wallet’s signers pay out.

There used to be a faster route. Optimistic minting let approved Minters issue tBTC before the sweep, with Guardians able to object. Threshold paused it on 18 May 2026 after what it describes as attempts to mint without depositing BTC, adding that “The attempts failed” and that no tBTC was issued. Threshold put the normal wait at about 6 to 7 hours after that, up from about 1.5 hours. In a forum post of July, the account mhluongo, writing for the Mezo dev team, gave 8 to 24 hours for bridging BTC to Mezo. The TBTCVault contract still reported optimistic minting as paused when we read it on 1 November 2026.

Fees are 0.2% on minting and 0.2% on redemption. The mint fee came back on 15 April 2026 under TIP-109, after a long holiday, and under TIP-093 any fee change has to wait 30 days. Holders can use tBTC on Ethereum, Arbitrum, Base, Solana and Starknet, according to Threshold, which relays a Q1 2026 Alea Research estimate of about 94% of supply on Ethereum.

The signing scheme is also on the list of things to change. Production uses threshold ECDSA built on the GG18 protocol, through a library Threshold maintains itself. Its team says the move to FROST and Schnorr signatures is “one of the team’s core priorities heading into Q4” 2026.

Why the tBTC bridge is paused, and what is known

The earliest report we read, from TokenPost, is stamped 00:17 UTC on 31 October and says Threshold stopped new BTC deposits, minting and new redemption requests in its app as a precaution during “security maintenance”. We could not read Threshold’s original post on X. LlamaRisk dates the pause to 31 October. The same TokenPost report says existing tBTC can still be held and transferred, and that users were told not to send BTC to deposit addresses generated before the pause. Redemptions already under way may be paid on Bitcoin, with final settlement on Ethereum on hold. Threshold’s September recap had said most of that month went into security hardening, but nothing we read connects that work to the pause.

On 31 October Threshold said, according to a The Defiant report carried by KuCoin, that tBTC remains fully backed, that “no funds have been lost” and that it had seen no unauthorized minting. It cited 4,320 BTC in bridge wallets against about 4,185 tBTC, and The Defiant counted 4,320.33 BTC itself across six of the 44 active wallets. We read the Ethereum supply: 4,185.155 tBTC on 1 November at about 19:38 UTC. We did not check the BTC side. Bridge functions return once “a full remediation is in place”, with no date given.

A separate report concerns the signer machines. A Phemex News item stamped 19:01 on 1 November, with no time zone, relays a post by The Defiant: T Network reported that all tBTC operators had either patched a Linux vulnerability or gone offline, and that it found no evidence the flaw was exploited against any tBTC node. This is one aggregator repeating one outlet. The Defiant’s article could not be opened, and we found no statement from Threshold on its blog, forum or GitHub, and no other outlet. The item names no vulnerability and no number of operators. Nothing we read says this flaw prompted the pause, so we treat the two as separate facts until Threshold says otherwise.

Who reacted: Aave, Compound and Morpho

Aave moved first. Its guardian multisig froze tBTC on Aave V3 Ethereum Core at 10:01 UTC on 31 October, on LlamaRisk’s recommendation, “until the incident is confirmed to be fully mitigated by Threshold”. The reserve was still frozen, not paused, when the Aave Pool configuration was read on 1 November at about 19:25 UTC. The Defiant’s report puts about $136.5 million of tBTC on that market. By LlamaRisk’s account, users can repay and withdraw, liquidations continue, and tBTC can no longer add borrowing power. The Base and Arbitrum reserves had been frozen in earlier rounds.

Gauntlet recommended that Compound set tBTC supply caps and borrow collateral factors to zero, with 101.2 tBTC, about $7.8 million, supplied across six Comet markets at 11:59 UTC on 31 October. Its reasoning applies to every holder: while redemptions are paused, the main mechanism that holds tBTC to BTC is unavailable. In a tBTC-backed market used by its Morpho USDC Core vault, Gauntlet withdrew about 65,000 USDC and about 640,000 USDC stayed supplied, according to The Defiant’s report.

Nothing we read reports a depeg. DefiLlama’s feed listed tBTC at $77,663 at 19:22 UTC and BTC at $77,740 at 19:27 UTC on 1 November, a gap of about 0.1%. The readings are five minutes apart, and a price feed can lag thin trading.

Who controls Threshold: signers, a 6-of-9 committee and a DAO

The operators move the bitcoin. The contracts around them answer to the Threshold Committee, a multisig wallet that needs 6 of 9 signatures. We read the relevant contracts on Ethereum on 1 November 2026:

• The Committee Safe (0x9F6e…2F5f) reports a threshold of 6 and nine owners.
• It owns the BridgeGovernance contract, whose parameter-change delay reads 48 hours, and the TBTCVault, which owns the tBTC token and reports a 24-hour governance delay.
• The Bridge is an upgradeable proxy. Its admin is owned by a timelock with a 24-hour minimum delay, and the Committee Safe holds the proposer role on that timelock.

So six signers can schedule a change to the core bridge, and users get about a day of notice. We did not read every role in the system. A custodian’s internal changes cannot be watched on-chain, and these can. But the phrase “trust-minimized” has limits here: part of the trust moves to nine keys, some held under pseudonyms.

Those keys changed in 2026. TIP-110, passed on Snapshot on 15 July with 555.5 million votes for, none against, and only four voters, cut the $649,000 approved for the Committee and part-time DAO roles to zero, apart from a Foundation director employed by the firm Marfire. The proposal names the new line-up by forum handle. Five members handle protocol upgrades, among them @maclane, the account of tLabs CEO MacLane Wilkison, the tLabs CTO, and @michwill, described in the text as the founder of Curve and Yield Basis and co-founder of NuCypher, which is Michael Egorov. Three others plus the Foundation director handle treasury matters. Each named account then posted its signer addresses on the forum, and on 1 November all nine owners of the Safe matched those addresses. That public record is our only source for who holds each key: we found no separate statement from either man or from Threshold’s website. On that record, the CEO and CTO of the company the DAO funds hold two of the nine keys, and the docs give the same Committee the power to veto a proposal.

Token votes follow a forum discussion and a Snapshot check. An on-chain proposal needs 0.25% of supply, or about 27.9 million T, and passes with a quorum of 1.5%. The docs still give a 10-day vote, but TIP-108 cut that to 7 days in December 2025, and the Governor contract reads 50,000 blocks, which matches. Execution waits for a two-day timelock. Our guide to governance tokens explains how such votes work in general. The last five TIP votes on Snapshot drew between four and 10 voters.

tLabs lives on the DAO’s money. TIP-104 gave it $2.85 million for its first year, passed with 803.5 million votes, and set two goals: 50 tBTC of annual revenue and 7,500 BTC of TVL. Wilkison waived his pay as CEO. In June 2026 it asked for $188,000 a month for June to August and described the next three months as “determinative for tBTC’s future direction”. We found no published scorecard or later budget. DefiLlama’s revenue for the last 12 months, $568,246, is about a quarter of what that monthly rate costs over a year, by our arithmetic, and supply of 4,185 tBTC on Ethereum sits well under the 7,500 target. We could not confirm the legal home of tLabs or of the Threshold Foundation.

The T token: 11.155 billion units, votes and a fee discount

The merger created 10 billion T: 4.5 billion for NU holders, 4.5 billion for KEEP holders and 1 billion for the DAO. Another 1.155 billion was minted for staking incentives, for a final 11,155,000,000. The contract’s total supply matched that figure when we read it on 1 November 2026, and there is no vesting calendar left: the docs call T fully diluted. CoinGecko shows no maximum supply, but emissions ended with a modified TIP-092 and the DAO’s choice to stop them by February 2025. Threshold’s supply API gives 253.3 million T in the DAO treasury.

That allowance covers the fee on about 0.5 BTC of activity. Alea, quoted in Threshold’s September recap, calls it “a fee discount for active users, not passive yield”. Even that may not last. On 25 June 2026 Wilkison proposed replacing the rebate contract with a DAO-run peg keeper with free minting and redemption, saying “staking uptake has been minimal”. Stakers objected, one arguing it would remove their only reward, and the mhluongo account backed the critics. We found no vote on it on Snapshot, and a Committee member wrote in September that the rebate remains active.

tBTC TVL and revenue: a small fee base

These are dollar figures, so part of the fall reflects the bitcoin price rather than coins leaving. In BTC terms, Threshold’s own TIP-104 gives a peak of 5,274 in November 2024, and the Ethereum supply now stands at about 4,185. DefiLlama counts $49,281 of fees over 30 days, $568,246 over 12 months and $2.84 million all-time, with nothing in the last days because of the pause. Revenue is all fees paid to the treasury after rebates. Alea, via Threshold, said BTC deposits and fee income both declined in Q3 2026.

tBTC vs WBTC and other wrapped bitcoin

tBTC is small next to the custodial leaders. The comparison below measures size, not safety, because the models carry different risks: WBTC depends on BitGo’s custody arrangement, while tBTC depends on its operators, its contracts and its Committee. Lombard, Solv and Function mix institutions, custody and other tools. cbBTC has no entry we could read on DefiLlama.

tBTC audits: what the reports cover

We opened 11 reports from Threshold’s audit page, listed in full on this page. Two more on that page, Zellic on Sui and Hashlock on Starknet, sit in a Notion page we did not open. The core bridge was reviewed by Least Authority in 2022, which found that a valid SPV proof could be built for a fake Bitcoin transaction, a flaw that could have allowed unbacked tBTC. It is marked resolved. MixBytes found one high-severity bug in cross-chain redemptions in September 2025, a call that reverted, also fixed. Certora’s CCIP review found no high or medium issues, and the copy linked by Threshold is still marked draft.

Three limits apply. Five of the 11 reports are by Thesis Defense, the auditing arm of Thesis, Inc., which those reports call “the venture studio behind tBTC”: the reviewer belongs to the group that started the product, and it is not an outside firm. The latest report is dated 12 December 2025, and none we saw covers the off-chain signing client or the planned FROST move. And the 2024 Base review warned that its “security-critical” relayer bot was out of scope. The Immunefi bug bounty pays 10% of the funds directly affected for critical smart contract bugs, from $10,000 up to $150,000.

Is tBTC safe? The incident record behind “zero losses”

Threshold’s brand guide speaks of a “track record of zero-loss security”, and its About page says “Five years live without incident”, while its May 2026 recap counts six years and calls the failed mints a zero loss event. That is the company’s own claim, and its numbers do not match each other. No loss of user funds is reported in the sources we read. The record does include incidents:

• 2020: the first tBTC paused deposits for 10 days after a flaw, then relaunched in September with a supply cap. Keep said no user funds were lost.
• May to June 2023: after a private warning about GG18 attacks, later known as BitForge and TSShock, Threshold paused new wallets and heartbeats on 11 May and says it was mitigated before disclosure.
• 30 September to 7 October 2023: crafted Bitcoin transactions made two wallets look busy and blocked all outstanding redemptions. Threshold also found that one address handled redemption maintenance, which it called “a single point of failure”.
• 18 May 2026: failed attempts to mint without BTC, followed by the pause of optimistic minting.
• 30-31 October 2026: the current pause, with the cause still unconfirmed.

Redemptions have stopped before: in 2023 for about a week, until a patch shipped on 6 October. How long the current pause runs is the open question.

Our risk rating and what to watch

We rate Threshold 4 out of 5 while the pause lasts. In its favour: a long record with no reported loss of user funds, open code, readable audits and on-chain delays anyone can check. Against it: no minting or new redemptions for now, an undisclosed cause in any source we could read, a closed list of about 20 operators, a six-signature path to upgrades, and fee income far below costs. Risk differs by token. A tBTC holder depends on the bridge reopening and on the secondary market until it does. A T holder has votes and a discount that the team itself proposed replacing. Coindaro makes no price forecast, and nothing here is investment advice.

Next dates are few. Threshold has given no restart time, and LlamaRisk says it will propose restoring the Aave reserve once operations are normal. tLabs’ October roadmap lists security hardening, FROST and Schnorr signing, segregated “reservations” with a working assumption of 20 basis points each for mint, redeem and yearly custody, and instant BTC swaps, all without dates.