Solana Foundation's stack pushes a single fast chain to its limits — the Firedancer client, built by Jump Crypto and now live on mainnet, targets over 1 million TPS, but high throughput has historically come with outages and rising validator hardware costs. Monad Labs (parallelized EVM L1, mainnet November 2025), Aptos Labs, and Mysten Labs' Sui pursue parallel execution and pipelining to raise throughput while trying to keep nodes affordable.

Celestia Labs separates data availability from execution, letting rollups post data cheaply to a dedicated DA layer rather than congesting an execution chain. This spreads the scaling burden across specialized layers, but introduces cross-layer complexity and new trust assumptions about which layer is the source of truth.

Offchain Labs (Arbitrum), OP Labs (Optimism/OP Stack), Polygon Labs, and Matter Labs (zkSync) push execution off Ethereum onto L2s that batch transactions and settle proofs to mainnet, keeping L1 security while cutting fees. The unsolved piece is that scaling L2s shifts decentralization questions to sequencers — which are often still centralized — rather than eliminating the trilemma.

OP Labs' Superchain (OP Stack chains sharing a settlement and messaging layer) and Polygon Labs' AggLayer aim to make chains within one ecosystem interoperate natively, so cross-chain transfers settle against shared security rather than an external multisig. The limitation is that this only solves interop within a family of chains, not across rival ecosystems.

zk-proof-based bridges (an area Matter Labs and the broader zk ecosystem push) let chain B cryptographically verify chain A's state rather than trusting a committee, shrinking the trust assumption. These are expensive to build per chain pair and still nascent for general message passing beyond simple asset transfers.

Rather than locking assets in a bridge contract, intent-based systems and liquidity networks have solvers fulfill a user's desired outcome on the destination chain, sidestepping the classic lock-and-mint honeypot. This reduces some bridge risk but concentrates trust and capital in solver networks instead.

Mysten Labs has signaled private transactions launching on Sui in 2026, an attempt to bring confidential transfers to a high-throughput general-purpose L1 rather than a niche privacy chain. The hard part is pairing confidentiality with auditability so the feature survives exchange listing and regulatory scrutiny.

zk technology — central to Matter Labs (zkSync) and Polygon Labs' zk roadmap — can in principle prove a transaction is valid (or a user is permitted) without revealing the underlying data, enabling compliance checks against hidden balances. Turning research-grade zk primitives into production privacy that regulators accept is still unproven.

Chainalysis, TRM Labs, and Elliptic build the surveillance side — tracing flows, screening sanctioned addresses, and giving banks comfort to engage. Their existence is exactly what privacy designs must contend with: any privacy feature that fully defeats tracing risks being treated as a mixer, so the equilibrium between the two camps is unresolved.

Hyperliquid runs a fully on-chain central limit order book on its own L1, an architecture that changes the MEV surface versus AMM-based DEXs by making ordering and matching part of the protocol itself. Purpose-built trading chains can design ordering rules directly, but whether they fully neutralize extraction is contested.

L2 teams like OP Labs and Offchain Labs control their own sequencers, which gives them a lever to implement fair-ordering or encrypted-mempool schemes that block front-running — but a centralized sequencer is also itself an MEV risk, so the fix and the threat are the same component until sequencer decentralization lands.

DeFi protocols such as Morpho, Avara (Aave), and Sky depend on price oracles for liquidations and collateral valuation; the standard mitigations are multiple independent feeds, time-weighted averages, and circuit breakers. These reduce single-feed manipulation but add latency and still failed in past exploits, so oracle robustness remains an active, unfinished problem.

Coinbase (as USDC's distribution partner and Base operator) and crypto-com push fully-reserved dollar stablecoins toward payments under the new GENIUS-Act regime. The open question is economic: the model is essentially a float business that is structurally short interest rates, and a bank-led consortium issuing rival coins could erode exchange distribution.

Because GENIUS-Act stablecoins can't pay yield, the demand for return is migrating to tokenized-Treasury tokens — Ondo Finance's USDY/OUSG and Sky's RWA-backed reserves — which accrue yield while behaving like dollars for non-US or institutional holders. This routes around the no-yield rule but lands squarely in securities-law territory.

MoonPay and Ramp Network are building stablecoin payment and conversion products to turn stablecoins from a trading instrument into spendable money. The unsolved piece is the last mile — merchant acceptance, cross-border licensing, and consumer reasons to hold a non-yielding dollar — which determines whether the payments thesis actually materializes.

Account abstraction (programmable smart-contract wallets) lets users add recovery guardians, spending limits, and session keys so a lost device isn't fatal — an area pushed across the Ethereum L2 ecosystem (OP Labs, Matter Labs, Offchain Labs). The challenge is that recovery schemes reintroduce trust in guardians and are still unfamiliar to mainstream users.

Multi-party computation splits a key into shares so no single device or party holds the whole key, the model behind Fireblocks' institutional custody and increasingly behind consumer embedded wallets. It removes the single-seed-phrase failure but shifts trust to the share-management scheme and the providers operating it.

Ledger (hardware) and Phantom (multi-chain app wallet) work the UX layer — clearer signing, scam detection, and recovery flows — to reduce user error without abandoning self-custody. These improve safety at the margin but have not eliminated the irreversibility problem that account abstraction tries to solve at the protocol level.

Uniswap Labs (the fee-switch debate), Avara (Aave's 'Aavenomics' buyback-and-distribute), and Sky route a share of protocol revenue back to the token, attempting to tie token value to real usage. The tension is that turning on fees can drive volume to cheaper competitors, so value capture and competitiveness pull against each other.

Hyperliquid (its own L1 plus the dominant on-chain perps venue) and Coinbase's Base show the monetize-the-volume path: capture the trading or settlement activity itself rather than per-transaction fees. Whether an app or chain can defend that volume as rivals replicate it is the open question.

OP Labs, Offchain Labs, and Polygon Labs look to sequencer fees, shared-infrastructure economics (Superchain/AggLayer), and selling rollup tooling to other projects as revenue lines beyond raw transaction fees. As Ethereum's roadmap keeps pushing L2 data costs down, it is unproven whether these lines scale faster than fees compress.