a16z Crypto says finance is one of the clearest use cases for blockchains, but also one of the most demanding. A shared system that records assets, ownership, and execution rules outside the control of any single counterparty may sound well suited to markets. That alone, the firm argues, does not mean blockchains are ready for the scale and operational complexity of global finance.
The article, authored by a16z Crypto and published by MarsBit in a Chinese translation credited to ChainCatcher, focuses on a narrower question: what capabilities blockchains still need if they are to serve as real financial infrastructure, and where progress has already been made.
Throughput improved, but market requirements go beyond speed
For years, debate over whether blockchains could support large-scale use centered on cost and throughput. Throughput measures how many transactions a chain can process per second. In finance, that matters: without enough capacity, routine actions such as placing, canceling, or updating orders become expensive and unreliable when network activity rises.
a16z Crypto says that constraint has eased materially over the past several years, especially after the last wave of infrastructure upgrades. Over a five-year period, total blockchain throughput across the industry increased by more than 100x, and some live systems can now process tens of thousands of transactions per second. The firm adds that this is likely not the ceiling, with more gains possible as stress testing continues and new architectures go live.
Still, throughput does not answer a more important market question: when exactly will a transaction be included in a block? For financial applications, the article says, that matters as much as raw speed.
Predictability: inclusion on time and ordering under known rules
The first capability is predictability. Many actions in financial markets are only valid within specific windows, so transactions must not only be processed; they must be included on time and handled under rules participants know in advance. A one-second delay may not matter much in everyday payments. In markets, the piece says, uncertainty measured in milliseconds can be enough to weaken trust in a trading venue.
The article uses an onchain order book to illustrate the point. If a trader’s cancel request does not take effect in time, someone else may trade against a quote that is no longer valid. Market makers price that risk into their quotes by widening spreads, which leaves all participants with worse execution.
In that framing, predictability has two parts:
- assurance that transactions can be included on time, described as censorship resistance; and
- assurance that transactions will be ordered under clear rules.
Resilience: participants need reliable paths to submit transactions
The second capability is resilience, which asks whether participants can reliably get transactions into the system at all. Even if ordering rules are well designed, a16z Crypto argues, the infrastructure problem remains if one actor still controls who can submit and execute trades.
The article offers an analogy to the New York Stock Exchange. Imagine, it says, that every second the exchange randomly selected one node from around the world among those participating in staking, and that node controlled the only channel through which all trades entered the market. Those nodes would hold enormous power over execution.
Ordering governs the sequence of orders once they are onchain. Resilience comes earlier: can participants submit orders in the first place, and do they have to depend on a single gateway or operator whose decisions can affect prices, risk, or outcomes?
That is why developers are shifting the target from 「transactions will eventually be included」 to 「transactions can be included in time」. The stricter goal, according to the article, is that a valid transaction arriving at the network in time should be included immediately, rather than receiving a softer promise that it will be handled soon. Financial markets need that level of certainty.
a16z Crypto researchers have proposed the concept and measurement of Strong Chain Quality. The core idea is to make sure a participant’s chance of entering the next round of execution does not depend on a single queue controlled by one operator. One approach described in the article is for a blockchain to reserve part of each block for transactions submitted through other network paths, creating multiple entry routes for valid transactions and reducing the control any one operator has over which trades are prioritized.
The aim is to turn block space into shared infrastructure with several points of access, not a single queue controlled by one entity. The article says this would require protocol designs more complex than many systems running today, but suggests the added complexity may be worth it if participants can trust that their trades will be processed properly even under heavy strain.
Ordering: markets need to know how trades will be sequenced
The third capability is ordering. Predictability depends not only on access, but also on how a venue decides sequence once transactions are pending. On many blockchains, transactions are not finalized one by one. They are grouped into batches, or blocks, which are then written to the ledger by builders or proposers.
In many current systems, builders or proposers have wide discretion over which transactions enter a block and in what order. That may simplify protocol design and may be acceptable for some applications. In financial markets, the article says, sequencing directly affects who gets priority, at what price a trade executes, and whether the process is fair.
It also creates information advantages. If one party controls the next block, that party may see trades before others and act on them. In crypto markets, this is often described as maximal extractable value, or MEV: value captured beyond ordinary block rewards by including, excluding, or reordering transactions.
The article points to the sandwich attack as a familiar example. A trader sees a pending transaction, trades ahead of it to move the price, then completes another trade immediately after the target order executes, profiting from the move. For institutional markets, a16z Crypto says, the issue is even simpler: whoever controls order flow may act on that information before everyone else.
That resembles weaknesses in existing market structure. In traditional markets, centralized intermediaries can profit from privileged access to order flow and trading intent. If onchain markets cannot provide strong guarantees around ordering and execution, the same information asymmetry may reappear on blockchains.
Protocol designers are now studying more explicit ordering rules. The article highlights two broad directions:
- deterministic rules based on priority fees; and
- rules set by specific venues according to their own applications.
The exact mechanisms remain an active area of research. The underlying principle is familiar from traditional finance: participants should be able to anticipate how their trades will be ordered and executed, much like price-time priority at established venues.
a16z Crypto stresses that inclusion and ordering must work together. If an auction bid or liquidation trade is selectively delayed before it reaches the ordering stage, a fair sequencing rule does not solve the problem. The reverse is also true. If participants cannot tell how a transaction will be handled after it is included, timely inclusion loses much of its value.
High throughput makes onchain finance possible, the article says. Only predictable inclusion and clear ordering rules make it usable when timing matters and markets are under pressure.
Privacy: protect trading intent before execution
The fourth capability is privacy. A transaction can reveal information a participant holds or what that participant plans to do next. The size and direction of an order may show whether a fund is building or reducing a position. If that information becomes public before the order is fully executed, other traders can move first and affect the price.
This is especially important onchain because many blockchains expose pending transactions before final confirmation. Even if post-trade information cannot be completely hidden, the article says markets still need privacy protections before execution.
The current goal for protocol designers is to keep transaction details hidden during the period when others could still use that information, while allowing more tailored privacy models after execution. Once the trade is complete, disclosure rules can change. A public market may publish a full record. A bank may disclose the relevant record only to counterparties and regulators. Others should still be able to verify that the trade followed the rules even if they cannot inspect every detail.
To get there, designers are exploring cryptographic tools including timelock encryption and threshold encryption. These approaches can keep transaction contents hidden until a preset deadline passes or a committee confirms that the transaction’s position in a block has been fixed. By the time the market can see the trade, front-running is no longer possible.
Transparency, in this view, does not require every participant to see every unexecuted order in real time. Onchain finance needs the same boundary: auditable after execution, confidential before execution.
What blockchains still need to become core financial infrastructure
a16z Crypto ends with a simple claim. If blockchains are to become core financial infrastructure, they must support a market participants can trust. Throughput is only the first test. Without enough capacity, onchain finance will struggle to move beyond niche use. Once speed becomes a baseline capability, the harder question is whether blockchains can meet the standards financial institutions already expect from market infrastructure.
Those standards, as laid out in the article, include predictable execution, clear and unambiguous priority rules, stable and robust transaction access, and protection for pre-trade information without giving up post-trade transparency.