On 27 January 2026, Citrea's mainnet posted its first zero-knowledge proof to Bitcoin and had it verified through a BitVM2 bridge contract — the first time in Bitcoin's seventeen-year history that a general-purpose rollup settled to L1 without trusting a federation. That single transaction is the punctuation mark on a two-year stretch that quietly rewrote what "Bitcoin DeFi" means.
This series is about that rewrite. Part 1 lays the map: why Bitcoin's base layer refuses to scale, what "L2" actually means on Bitcoin (it is not one thing), and where each option sits on the only axis that matters — how much you have to trust someone other than Bitcoin itself.
Why Bitcoin's L1 stays small on purpose
Bitcoin targets a ten-minute block time and caps each block at roughly 4 million weight units, which works out to an effective base size around 1 MB. That ceiling is not an oversight. It is the price Bitcoin pays so that more than 23,000 globally distributed full nodes can validate every block on consumer hardware. Lift the cap and you start pricing out the home validator; price out the home validator and you have a different system.
Bitcoin Script reinforces the same conservatism. It is intentionally non-Turing-complete: no loops, no global mutable state, no unbounded execution. Every output script terminates in a knowable number of operations, which is what lets a node in Buenos Aires reach the same conclusion as a node in Helsinki without coordinating.
The soft-fork track tells the rest of the story. Taproot — BIPs 340, 341, and 342 — activated at block 709,632 on 14 November 2021. It is still the most recent consensus change to Bitcoin. BIP-119 (OP_CHECKTEMPLATEVERIFY) and BIP-118 (SIGHASH_ANYPREVOUT) have been drafts for years and remain drafts in May 2026. The lesson developers have absorbed is brutal but clear: if you want programmability on Bitcoin's timeline, you build it off-chain and anchor.
The trust spectrum, from custodial to (almost) trustless
Every "Bitcoin L2" answers the same question — who controls your BTC while it is being used elsewhere? — with a different trust model. Five tiers, weakest to strongest:
| Tier | Example | Who can take your BTC |
|---|---|---|
| Custodial wrapper | WBTC, cbBTC | A single custodian |
| Federated peg sidechain | Liquid, Rootstock | 2/3 of a fixed federation |
| Threshold-signer rollup | Stacks / sBTC | A supermajority of elected signers |
| BitVM2 optimistic bridge | Bitlayer, Clementine | Only if every honest watcher fails |
| BitVM-verified validity rollup | Citrea, Alpen | Only if Bitcoin itself fails |
Custodial wrappers like WBTC mint an ERC-20 IOU against BTC held by BitGo. Convenient, liquid, and a single subpoena away from being frozen. WBTC's market cap still sits near $8.9-9B in May 2026 — bigger than every trust-minimised option combined by an order of magnitude.
Federated pegs are a step up. Liquid Network launched 27 October 2018 with a 2/3-of-functionaries signing federation; Rootstock launched in January 2018 and is now merge-mined by roughly 80% of Bitcoin's hashrate. You are still trusting a fixed group, but the group is larger and the rules are public.
Threshold-signer rollups like Stacks tighten this further. sBTC's 14 community-elected signers must hit a 70% (10/14) approval threshold to move pegged BTC, and the signer set rotates. Better than a federation, still not Bitcoin-equivalent.
BitVM2 bridges are the qualitative leap. They use optimistic fraud proofs verified directly on Bitcoin: a single honest watcher, anywhere on Earth, can challenge a fraudulent withdrawal and the Bitcoin chain itself adjudicates. Validity rollups go further still, posting zero-knowledge proofs that a BitVM2 wrapper verifies on L1.
The practical takeaway is that "Bitcoin L2" tells you almost nothing on its own. The trust tier tells you everything.
What changed in 2024-2026
For most of Bitcoin's history, the trust-minimised end of that spectrum was theoretical. Three things happened in eighteen months that changed it.
BitVM and BitVM2. Robin Linus of ZeroSync published the original BitVM paper in October 2023, showing that arbitrary computation could be verified on Bitcoin using existing opcodes and challenge-response games. BitVM2 followed on 15 August 2024 with practical bridge constructions. No soft fork required.
Stacks Nakamoto and sBTC. The Nakamoto upgrade activated on 29 October 2024 at Bitcoin block 867,867, giving Stacks fast finality anchored to Bitcoin. sBTC mainnet deposits opened 16-17 December 2024; permissionless withdrawals went live on 30 April 2025. For the first time, a major BTC-pegged asset operated under a rotating, on-chain-elected signer set rather than a fixed multisig.
Babylon. Phase 1 of Babylon's self-custodial BTC staking launched on mainnet 22 August 2024. Phase 2 — the Babylon Genesis chain that lets stakers secure other proof-of-stake networks — went live in April 2025. By 2026 it had attracted roughly $3-3.4B in BTC, all of it staked from users' own wallets without ever leaving Bitcoin.
Production BitVM bridges. Bitlayer's BitVM Bridge launched on mainnet 16 July 2025, the first production BitVM2 deployment. Citrea followed on 27 January 2026 with a full ZK-rollup using the Clementine BitVM-based bridge.
Before this window, BTCFi meant give your coins to a custodian. After it, BTCFi means pick your trust tier.
BTCFi by the numbers in Q1 2026
The Block's Q1 2026 readout pegs total BTCFi TVL at roughly 91,332 BTC — about 0.46% of circulating Bitcoin supply. That is small in percentage terms and large in absolute terms, and the distribution tells the real story:
- WBTC: ~$8.9B. Custodial. Still the dominant pool of "BTC" on smart-contract chains.
- Babylon: ~$3-3.4B in BTC restaked across Bitcoin Secured Networks, all self-custodial.
- Stacks: closed Q1 2026 with ~$437M in sBTC TVL after a Q1 peak near $545M; deployed Stacks DeFi capital sat around $121M.
- Rootstock: ~$272M TVL with ~$278M of canonical BTC bridged in.
- BitVM-class bridges (Bitlayer, Citrea, others): still small in TVL but growing into the gap.
The headline: custodial BTC on Ethereum is still an order of magnitude larger than every trust-minimised alternative combined. The trend is the opposite — every quarter since sBTC launched, the trust-minimised share has crept up — but anyone telling you BTCFi has "gone trustless" is selling the future as the present.
Why holders should care now (and why some shouldn't)
If your thesis on Bitcoin is "sound money, hold and forget," most of this is noise. Holding BTC in self-custody and waiting out the halving cycles is a complete strategy and always has been. Nothing in BTCFi changes that.
If your thesis is "BTC as productive capital," the L2 trust model is the whole investment. Three things to internalise:
- Self-custody yield is finally real. Babylon lets you sign a staking transaction from your own wallet, with your coins locked by a Bitcoin script you can audit. No custodian, no IOU.
- Settlement guarantees vary by orders of magnitude. A federated peg can rug if 2/3 of functionaries collude. A BitVM bridge requires Bitcoin itself to break. These are not the same risk.
- Bridge risk dwarfs smart-contract risk in historical losses. Pick an L2 the way you pick a custodian — read the trust assumptions before you read the APY.
A useful exercise: before bridging any BTC, write down who can take it and under what conditions. If the answer is longer than two sentences, you do not understand the system you are using yet.
Series roadmap
This was the map. The next four parts walk each tier in detail.
- Part 2 — BitVM2-era rollups: how Citrea and Alpen actually generate validity proofs, what the Clementine bridge does on every withdrawal, and where the optimistic-vs-validity tradeoff lands.
- Part 3 — Stacks Nakamoto and sBTC's threshold-signer model: signer election, the 10/14 quorum, and what happens at peg-out.
- Part 4 — Babylon and Rootstock: the self-custodial staking design and the longest-lived merge-mined sidechain.
- Part 5 — A 2026 operational guide: choosing an L2, sizing exposure, and bridging without losing keys or sleep.
Part 2 picks up where Citrea's first proof landed: the moment Bitcoin verified a rollup it had never met.
