Raw transaction speed is becoming less important as a criterion for judging whether a blockchain can support heavy trading, according to a blog post published by venture capital firm a16z crypto on September 3. The more relevant questions are whether transactions arrive on time and whether they execute in a predictable order.

For banks, asset issuers, and market makers moving into the on-chain sector, the distinction matters. A network capable of processing thousands of transactions per second is still a poor choice for trading if a single player can delay or reshuffle those transactions.

Why throughput stopped being the differentiator

Blockchain capacity has improved dramatically. In its 2025 State of Crypto report, a16z reported that combined throughput across major networks has increased more than 100-fold over the past five years, rising from fewer than 25 transactions per second to over 3,400 transactions per second. The September blog post added that some production systems can already process tens of thousands of transactions per second.

But raw speed does not tell traders when a specific transaction will be processed. A payment can tolerate a one-second delay; markets cannot, particularly during periods of rapid price movement. In traditional equities markets, this is why rules such as best-execution requirements and fair-order-handling obligations exist — protections that on-chain markets have yet to fully replicate.

How a fast chain can still execute badly

a16z illustrates the problem with on-chain order books. Suppose the Federal Reserve makes an announcement that causes prices to spike. Market makers rush to cancel stale quotes and post new ones. If those cancellations are not delivered on time, arbitrageurs can exploit the outdated prices for profit before the market makers' new orders land, leaving the market makers with losses.

The firm examined this issue in its earlier research on predictable onchain execution. The risk is that market makers respond by widening spreads, harming all participants through worse prices and reduced liquidity.

a16z defines predictability through two criteria: valid transactions must be included in the process, and they must be ordered according to rules that participants can understand in advance. The firm's researchers have also proposed Strong Chain Quality, a design that guarantees stakeholders access to portions of each block, limiting the influence of any single block proposer or other participant.

The reordering risk regulators already flagged

The issue is closely tied to maximal extractable value (MEV), in which those controlling transaction ordering profit by reordering, including, or excluding transactions. A well-known example is the sandwich attack, where an attacker places a transaction immediately before and after an innocent user's transaction to profit from the resulting price movement.

Regulators have already raised alarms. In June 2022, the Bank for International Settlements cautioned that blockchain validators might use transaction sequencing to engage in front-running and sandwich trades that would not be permitted in traditional markets. In December 2023, IOSCO issued nine policy recommendations on decentralized finance focused on market integrity and investor protection.

A paper on fair ordering written in September 2026 reports that unfair transaction ordering and MEV have extracted more than $686 million from Ethereum users. The paper cites a previous MEV study for that cumulative figure rather than presenting it as the 2026 loss total.

What it means for institutions coming onchain

The stakes grow as more capital moves onchain. Cryptopolitan previously reported that malicious sandwich activity had shrunk to a small share of Solana blockspace activity as the network adopted more efficient ordering, private transaction routing, and confidential execution.

"Malicious extraction now represents a very small fraction of blockspace activity, while the majority of transaction ordering value reflects legitimate competition for inclusion and speed."

— Lucas Bruder, co-founder and CEO of Jito Labs

a16z frames the argument in broader terms. Only chains that can deliver fair and efficient execution under stress are likely to attract institutional order flow and, in turn, deep liquidity. The necessary protocols are more complex than the technologies currently in use, the firm argues, and best-execution mechanisms for ordering remain a "work in progress."