There were a lot of Drive Chain supporters at Tabconf this year, including Paul Sztorc himself. I am still a little fuzzy on some details, but I came home and looked at the very extensive https://www.drivechain.info/ and got most of my questions answered (not that the drivechainers won me over, but just that I think I now understand what they are proposing).
Seeing as everyone is arguing about filters and the price of bitcoin is "crashing," @BitcoinErrorLog has come out swinging against drivechains with this nice little piece:
Drivechains do not increase Bitcoin throughput. They move activity off chain while keeping deposits and withdrawals on chain. They substitute the peer-level trust of direct multisig constructions with miner control over exits.Drivechains would also introduce new incentive and governance problems, while repeating the same fragmentation and risk-pooling required in other layered systems.These factors ultimately undermine any ability to provide a credible exit to millions of users.1. Failure to provide trustless scaling due to onchain congestion
All value enters and leaves drivechains via mainchain transactions, just like any layer. Withdrawals are serialized “bundles” with a ~26,300-block (~6-month) voting window requiring ~13,150 ACKs. Only one withdrawal can succeed at a time per sidechain. Each bundle can pay out up to 6,000 outputs. This creates a hard cap on exit throughput and a natural queue during churn or stress. That throughput is shared with all other layers, drivechains, and L1 bitcoin use cases. There is no credible exit for millions of users.Proponents will answer this with “most users will not withdraw.” The FAQ states “only specialists will use the BIP300 withdrawal,” but this admits dependence on custodial or specialist intermediaries and confirms that a universal user exit is not a feature.2. Trust is reassigned to miners
Withdrawal correctness is not enforced by user keys or fraud proofs. It depends on miner signaling over many blocks; a majority can delay, censor, or coordinate to misdirect funds within the voting window. The longer the window, the easier coordination becomes. BIP-300 explicitly describes this as “hashrate escrow,” effectively a 13,150-of-26,300 miner “multisig.” Users cannot force redemption within a bound.The usual counterclaim is “miners will not burn future fees.” That is a hope, not a guarantee. Payoffs, regulation, or sidechain-specific politics will distort incentives.3. Interoperability is still fragmented
Each sidechain is a separate ruleset. Payments across sidechains will require swaps, bridges, trusted providers, or other rails. BIP-301 also allows only one accepted BMM request per L1 block per sidechain, which does nothing to unify payment protocols or remove bridging. Whether there are 14, 40, or 400 sidechains, liquidity and protocols are split. This is the same fragmentation seen in hub or mint models, with added miner-gated exits.Each drivechain will also come with multiple incompatible payment protocols, failing the designers own appeal to the network effects of payment protocols. Why? Because there is no bitcoin in any of the drivechains, and no actual peg.4. Arbitrage parity does not hold under delay and risk
Redemption is slow and uncertain, and units can not trade at 1:1 with real BTC.A minimal pricing model makes this obvious:Value of 1 sidechain BTC ≈ p * e^(-rT) - kp: probability the withdrawal actually pays T: expected time to redemption r: cost of capital k: risk premium for liquidity, chain bugs, censorship, and governance gamesWith T measured in months and p less than 1, parity fails. Markets price time and uncertainty. Discounts will persist.Observable falsifiers (publish per sidechain): median realized T, exit backlog depth vs bundle capacity, observed failure/censorship incidents (proxy for p), and the executable basis between DC-BTC and BTC after fees and time discount.Expected lifecycle for a typical drivechain:Launch: thin liquidity, immediate basis below 1 BTC. Adoption: custodians/market makers bridge internally; on-platform IOUs drift toward par, on-chain DC-BTC stays discounted. Stress: exit queue extends (T↑), miner hesitation (p↓), discount widens; specialists gate or charge spreads. Consolidation/abandonment: either users accept spreads or the chain becomes a zombie with stranded dust. Policy pressure: larger chains attract rent extraction and soft censorship (p↓, k↑).5. Miner incentives are distorted on L1
Blind merged mining and withdrawal control create new revenue levers: delay, censor, or prioritize certain chains. Large sidechains increase the legal and political surface around miners. If miners can steer or seize funds, courts and regulators will try to make them do so. This couples miner revenue and governance to off-chain politics.A simple deviation condition: Let F = PV of expected future BMM fees from a sidechain over horizon H; S = one-shot value at risk in the peg; C = coordination, reorg, legal, and reputational cost.A miner coalition deviates when S − C > F. Popular chains (S↑), bear markets (F↓), jurisdictional capture (C↓), or large exit queues (users willing to pay) all push toward deviation. None of this requires universal theft; selective delay and rent extraction are sufficient. Technical analyses also point to coinbase signaling mechanics and equivocation/reorg pitfalls that add sharp edges.6. Altcoins are cleaner for experimentation
Altchains isolate risk. If an altchain fails, Bitcoin miners and Bitcoin consensus are unaffected. Their assets price risk openly instead of hiding it in exit queues, spreads, and governance frictions. Interop needs swaps either way; adding miner custody does not remove that complexity and adds systemic coupling to Bitcoin’s miners. Independent reviews have also noted that merged-mining cost and incentive claims can introduce centralization pressure.