Sure,

For the entire time that Bitcoin has existed, there's been two primary ways of transacting off-chain in a way that is in some way tethered to the Bitcoin blockchain: the Lightning Network and sidechains.

Lightning Network

The Lightning Network is only for payments and doesn't host apps—for this reason and for other reasons I won't go into here (UX, inbound liquidity, clunkiness) the network usage and the kind of stats we usually use to measure adoption (such as TVL) is vanishingly small compared to Ethereum L2s (less than 5,000 BTC, i.e. $300m).

Sidechains (Rootstock, Liquid)

Sidechains have not historically been called "L2s", although this unfortunately seems to be changing as of late. The idea of a sidechain is very old, and it most commonly refers to a group of signers that control funds on one chain in a mulitisig fashion such that L1 BTC in the multisig are redeemable 1:1 for tokenized BTC on the sidechain, using trusted federations.

For many years, Liquid and Rootstock have been the primary examples of such Bitcoin sidechains. Liquid includes some marginal extensions on the scripting language compared to Bitcoin, and Rootstock is a Bitcoin sidechain that runs an EVM fork.

Stacks (xBTC, sBTC, Nakamoto sBTC)

Stacks, although slightly different from Liquid and Rootstock (Stacks contracts had the ability to regulate their outcome contingent on things that happened on the Bitcoin L1) has also fit into the sidechain category, especially since xBTC (Stacks' representation of bitcoin) was also controlled by a trusted federation. Because Stacks and Stacks contracts listened to the Bitcoin mainnet as a part of its protocol there was some online arguing about whether Stacks was "more than a sidechain", but ultimately its founder did not make suggestions that Stacks was a Bitcoin L2.

Around December 2022, Stacks felt that they had invented a new way of bridging in Bitcoin that would provide it with cryptoeconomic security. This tokenized BTC was called sBTC which would replace xBTC. The paper was titled sBTC: Design of a Trustless Two-way Peg for Bitcoin and argued that Stacks had access to a secure STX/BTC price oracle through its PoX consensus protocol.

The idea was flawed and as I pointed out here and here, it would not work securely without introducing centralized oracles in a Chainlink-fashion. I did not keep close track of sBTC after this realization, but I did meet one of the core protocol designers at Token2049 last year who confirmed to me that this was the case (that sBTC could not function with cryptoeconomic security alone).

Stacks is currently undergoing an upgrade to the latest version called Stacks Nakamoto. Stacks Nakamoto appears to include a marginal change in how finality in Stacks is handled. It is unclear to me whether this is actually an upgrade to Stacks or a hotfix for how Bitcoin miners have been able to abuse Stacks' vulnerable block production.

Stacks Nakamoto does not appear to introduce any fundamental changes to how sBTC operates, and I can't think of any reason why the previous issues with sBTC would be resolved by a finality change. It does appear that Stack have started to self-label as a "Bitcoin L2", although the whitepaper admits that Nakamoto does nothing to address the issue of Stacks lacking the ability for users to unilaterally withdraw their funds from their system (the core definition of an "L2", atleast as it has been used historically).

I'm happy to investigate Stacks Nakamoto and the consequences for sBTC further in a separate invocation.

tBTC

tBTC (v1) was a project that took a very intelligent view on the Bitcoin sidechain dilemma. It solved two problems at once:

• Instead of trying to spin up a whole new network as a Bitcoin sidechain, it simply bridged out the bitcoin asset into the vibrant Ethereum chain
• It was decentralized, as good as trustless and cryptoeconomically secure

However, being "trustless and cryptoeconomically secure" does not come cheap. tBTC was overcollateralized with 150% in crypto-collateral (ETH) on the Ethereum-side. This made tBTC so inefficient compared to WBTC that its advantages in decentralization and trustlessness did not outweigh its challenges in search for product-market-fit.

The new version of tBTC (v2) does not use ETH as collateral anymore. Instead tBTC v2 is a large multisig of signers who stake their own token called "T" (but to my understanding tBTC is not redeemable for the appropriate amount of T, as it worked in previous versions).

To answer your question around a comparison between sBTC and tBTC, you can assume that both ultimately rely on trust. tBTC is meant to work in any network while sBTC is native to just Stacks. sBTC is also less proven and its reliance on trusted oracles is not clearly defined.

Regarding Mezo: Mezo, to my understanding is an EVM sidechain. They're using tBTC as the native asset and is built by the same competent team. They're currently focusing on a rapid GTM strategy, so in terms of technology, they have not shared anything exotic. It's an EVM sidechain for now. They call it "Bitcoin-aligned", like Polygon's PoS chain was "Ethereum-aligned" only, before they started to experiment with actual rollup technology.

New class of Bitcoin L2s

There has not been much interest in Bitcoin sidechains historically. If you think about it, there's little reason to use a Bitcoin sidechain instead of Ethereum. The Ethereum ecossytem has more apps, greater network effects and superior infrastructure. Since the trust assumptions of BTC inside a Bitcoin sidechain is that of a multisig, the nuances between Bitcoin sidechains and "Wrapped Bitcoin on Ethereum" are not that large.

All of that changed in 2023, for two reasons:

• Ordinals
• BitVM

Ordinals and inscriptions introduced both social and technical changes to Bitcoin. There's now a native community of NFT and token-enjoyers on Bitcoin that did not exist before.

The technical change is the way in which inscriptions pack arbitrary data into the Bitcoin blockchain which I detail closely in this talk. By leveraging SegWit and Taproot, you can now propagate 400kb-sized data blobs that will be picked up by any Bitcoin miner, or 4MB if you use a service like Slipstream, wihout having to split up these blobs over multiple transaction inputs as previous data packing schemes required.

You can leverage this ability to create arbitrary metaprotocols ontop of Bitcoin. Arbitrary metaprotocols also go under the name of "sovereign rollups". The Omni Layer, which Tether used to first create USDT tokens on Bitcoin many years ago, is one example of a sovereign rollup. The Ordinals protocol itself is a sovereign rollup.

Note that these aren't Layer 2s though: a Layer 2 would ideally find some way to aggregate transaction data rather than just dumping a whole blockchain inside another blockchain. Examples of teams building sovereign rollups are Bison Labs, Alpen Labs and Chainway.

BitVM is perhaps the largest reason why there's a renewed interest in Bitcoin L2s. It is believed that BitVM has the ability to optimistically resolve arbitrary computation inside Bitcoin Script itself, opening up a window for rollup-like constructions to exist natively on Bitcoin. Almost all current L2 participants are exploring BitVM today in some fashion—including Stacks, Chainway, Alpen Labs and even Rootstock. At Taproot Wizards we have some reservations about BitVM-style Bitcoin L2s which you can gauge by watching these video clips and reading Shinobi's summary.

I'll note that this is a fluid conversation, and people like Sergio Lerner from the Rootstock team has insisted that "BitVMX" which he proposes does not have these issues and I'm currently exploring the new papers that have come out in conjunction with this.