Bitcoin's Layer 2 Wars: The Landscape Nobody Is Winning
There’s a weird split happening in Bitcoin’s Layer 2 ecosystem right now. On one side, the Lightning Network (Bitcoin’s oldest and most boring L2) just hit an all-time high capacity of 5,637 BTC in December 2025, with transaction volumes surging 266% year-over-year. On the other side, a crowded field of DeFi-focused Bitcoin L2s is fighting over a shrinking pie: total Bitcoin DeFi TVL actually declined 10% in late 2025, dropping from 101,721 BTC to 91,332 BTC.
That’s 0.46% of all Bitcoin in circulation. A rounding error.
So what’s going on? Dozens of projects (Stacks, Merlin Chain, Hemi, Rootstock, the new Arkade beta) are all racing to be the Ethereum L2 for Bitcoin. EVM compatibility, ZK-rollups, smart contracts, DeFi primitives. And yet the total addressable market keeps shrinking while Lightning quietly dominates the one thing Bitcoin L2s were supposed to solve: actually moving Bitcoin.
The Lightning success story nobody talks about
Lightning is not sexy. It’s been around since 2018. It doesn’t have airdrops or token incentives or VC-backed TVL farming schemes. It just… works.
- 5,637 BTC capacity (ATH, reversing a year-long decline)
- 8+ million monthly transactions
- 266% year-over-year volume growth
- 99%+ payment success rate
- Sub-0.5 second settlement
That’s real usage. Not “TVL locked in a bridge waiting for governance tokens to vest” usage. Actual peer-to-peer payments, cross-border remittances, merchant adoption. CoinLaw’s 2026 Lightning statistics show the network handling meaningful volume, and the capacity chart turning upward after consolidation suggests bigger players are taking it seriously.
The architectural difference matters. Lightning doesn’t rely on bridges. It’s native Bitcoin, locked in non-custodial 2-of-2 multisig channels. The security model is simple: if your counterparty tries to cheat, you can prove it on-chain and take their funds. No wrapped tokens. No federation trust. No threshold signers.
But Lightning doesn’t do DeFi. It does payments. The projects trying to do DeFi on Bitcoin face different challenges.
The fragmented DeFi L2 landscape
Here’s where the money is (or was) in Bitcoin DeFi, according to data from The Block’s 2026 Layer 2 Outlook:
| Network | Type | Approx. TVL | Status |
|---|---|---|---|
| Merlin Chain | ZK-Rollup | ~$1.7B | 150+ dApps, EVM-compatible |
| Hemi | Modular L2 | ~$1.2B | 90+ protocols, institutional-backed |
| Stacks | PoX smart contract layer | ~$250M+ target | sBTC growing, Nakamoto upgrade live |
| Rootstock (RSK) | Merge-mined sidechain | ~$185M | EVM-compatible, institutional push |
| Liquid | Federated sidechain | Niche | Exchange settlement focus |
| Arkade (Ark) | VTXO-based | Negligible | Public beta since Oct 2025 |
Each network has a different trust model and developer ecosystem. Merlin uses ZK-proofs but relies on bridge security. Stacks uses Proof of Transfer and a threshold signature scheme for its Bitcoin peg (sBTC). Rootstock is merge-mined with Bitcoin but uses a federated bridge. Hemi is modular but newer and less battle-tested.
None of them are interoperable. Transaction volumes across all these networks combined fall well short of Lightning’s. And collectively, they’re losing ground.
Why the TVL is shrinking
The Block’s December 2025 data is blunt: Bitcoin DeFi TVL dropped 10% from 101,721 BTC to 91,332 BTC in late 2025. That’s the opposite of what you’d expect if Bitcoin DeFi were “taking off.”
A few reasons this might be happening:
1. Bridge risk is real. Most Bitcoin L2 TVL flows through bridges with varying trust assumptions. ChainsCore Labs warned in June 2025 that Bitcoin L2s face an “unseen infrastructure crisis”: Bitcoin has no native data availability layer like Ethereum’s EIP-4844. L2s are using OP_RETURN, Taproot scripts, or off-chain side channels, creating “fragile, capacity-constrained data pipelines.”
That’s technical jargon for: if the bridge breaks, your Bitcoin is gone.
2. Developer ecosystems are small. Stacks uses Clarity (a new smart contract language designed for Bitcoin). Rootstock and Merlin use EVM, which helps, but the Ethereum developer talent pool is already stretched thin across dozens of Ethereum L2s. Why would they migrate to a smaller, riskier Bitcoin L2 ecosystem?
The Block’s report found no evidence of Ethereum-to-Bitcoin developer migration. The migration trend in 2025 was within Ethereum’s ecosystem: L1 projects like Celo and Lisk moved to OP Stack L2s. Bitcoin L2s are growing their own communities, but slowly.
3. Incentive farming distorts TVL. A lot of the TVL in Merlin and Hemi likely came from token incentive programs, not organic DeFi usage. When those programs wind down, so does the capital.
4. Bitcoin holders don’t want DeFi risk. This is speculative, but worth considering: people who hold Bitcoin often do so because they don’t trust complex financial systems. Bridging BTC to a sidechain to farm yield on a lending protocol is the opposite of that ethos.
What’s actually being built
Despite the shrinking TVL, there’s real development happening. Let’s break down what each major L2 is trying to do.
Stacks: the institutional DeFi play
Stacks had a big 2025. The Nakamoto upgrade went live, giving Stacks transactions Bitcoin finality. sBTC (a decentralized Bitcoin peg using threshold cryptography) launched and hit 5,000 BTC by May 2025, with a target of 21,000 BTC.
The game-changer: Fireblocks integration in February 2026. Fireblocks connects 2,400+ institutional clients. Now they can mint sBTC, interact with Stacks DeFi protocols (Zest for lending, ALEX for swaps), and custody everything in a compliant institutional-grade platform.
If Bitcoin DeFi goes institutional, Stacks is positioned for it. But “if” is doing a lot of work in that sentence.
Merlin Chain: the TVL leader with bridge risk
Merlin Chain is the biggest Bitcoin L2 by TVL (~$1.7B as of August 2025). It’s a ZK-rollup, EVM-compatible, with 150+ dApps including MerlinSwap (a DEX).
The problem: all that TVL came through a bridge. If the bridge security fails, the whole thing collapses. ZK-proofs provide computational integrity, but they don’t secure the bridge itself. That’s still a trust point.
Arkade: the new architecture
Arkade launched its public beta in October 2025, backed by Draper Associates and designed by the team behind Ark Protocol. It’s architecturally interesting: instead of using payment channels like Lightning, Arkade virtualizes Bitcoin’s UTXO model using “Virtual Transaction Outputs” (VTXOs), short-lived notes that expire after ~4 weeks.
Bitcoin Magazine called it “Bitcoin’s most advanced native Layer 2 since the Lightning Network’s debut nearly a decade ago.” The promise: no channels to manage, no liquidity routing issues, but the same non-custodial security guarantees as Lightning.
The catch: it’s brand new. No TVL data. No battle-tested track record. And it relies on a central server (Ark Service Provider) for coordination, not for security, but for liveness. If the server goes down, so does the network (though your funds are safe and you can unilaterally exit to on-chain Bitcoin).
Whether this architecture can carve out a niche remains to be seen. It’s too early to call it a Lightning competitor.
Rootstock: the quiet veteran
Rootstock has been around since 2018. It’s a merge-mined sidechain (inheriting ~60% of Bitcoin’s hashrate) with EVM compatibility and a federated bridge called Powpeg.
TVL is modest (~$185M as of April 2025), and Messari reported “modest growth” in Q3 2025. But Rootstock is methodical. It’s launching institutional vault strategies, targeting idle Bitcoin sitting in cold storage, and pushing the Lovell upgrade to deepen decentralization.
It’s not sexy. But it’s been quietly functional for years, which counts for something in crypto.
The bridge problem nobody wants to talk about
Here’s the uncomfortable truth: most Bitcoin L2s are only as secure as their bridges. And Bitcoin bridges are hard.
Ethereum has native data availability (DA) through EIP-4844, which gives rollups a place to post their state data cheaply and verifiably. Bitcoin doesn’t have that. So Bitcoin L2s improvise:
- OP_RETURN can store 80 bytes per transaction, not nearly enough for a rollup.
- Taproot scripts can hold more, but they’re expensive and not designed for this.
- Off-chain DA layers introduce new trust assumptions.
ChainsCore Labs warned this creates “fragile, capacity-constrained data pipelines.” Translation: if a Bitcoin L2 rollup gets popular and tries to post state data to Bitcoin at scale, it’s going to hit a wall.
This is an unsolved infrastructure problem. Lightning, which doesn’t need a bridge or a DA layer, remains the most robust Bitcoin L2 by a wide margin.
What happens in 2026?
The question worth asking: who is Bitcoin DeFi for?
If you’re a Bitcoin holder who wants to earn yield, you can:
- Lend BTC on a CeFi platform (custodial risk)
- Bridge to Ethereum and use WBTC (bridge risk, but mature ecosystem)
- Bridge to a Bitcoin L2 like Stacks or Merlin (bridge risk, smaller ecosystem)
The value proposition for Bitcoin L2 DeFi isn’t clear. WBTC alone has ~$14 billion in TVL on Ethereum, more than double the entire Bitcoin L2 ecosystem combined. For DeFi functionality, Ethereum already does it better.
What Bitcoin L2s could excel at is payments and programmability without bridging. That’s what Lightning does well. That’s what Arkade is trying to do. But that’s not where the VC money or the TVL hype has been flowing.
The Bitcoin L2 wars might not produce a winner. Lightning owns payments. Ethereum owns DeFi. And the middle ground (programmable Bitcoin that doesn’t require bridging to Ethereum) might be a smaller market than the hype suggests.
That doesn’t mean nothing interesting will happen. Stacks + Fireblocks could unlock real institutional Bitcoin DeFi. Arkade’s VTXO model could prove compelling for specific use cases. Rootstock’s slow, methodical approach might outlast the airdrop farmers.
But if you’re waiting for a “Bitcoin L2 summer” like Ethereum had in 2021? The data says it’s not coming. The TVL is shrinking. The developer migration isn’t happening. And Lightning, boring and unglamorous as it is, is the only L2 that’s unambiguously winning.
Sources
Sources: The Block, Bitcoin Magazine, CoinLaw, BingX Academy, Gate.io, Hemi, Parallel Research, Messari, Stacks, CoinMarketCap, StackingDAO, ChainsCore Labs, Bitcoin Optech. Data as of February 2026.