BNB Chain outlines plans for a high-performance EVM layer-1, with CZ backing BNB as its sole token

BNB Chain wants to build a second layer-1 blockchain, and it is aiming for speeds normally reserved for centralized exchanges.

Updated BNB Chain documentation describes a high-performance, EVM-compatible network provisionally called NewL1. It is designed for latency-sensitive work such as high-frequency trading and AI-driven activity.

Founder CZ has weighed in with a reassurance aimed squarely at existing holders. BNB, he affirmed, will remain the only token for gas fees, staking and governance across the BNB Chain networks.

What the docs actually describe

The documentation was updated on October 22, 2026. It lays out a set of performance targets that read like a wish list for anyone who has ever watched an order fail during a volatile market.

The headline target is a 200 ms block interval. That means the chain aims to produce a new block five times per second.

Pre-confirmations are targeted at approximately 20 ms. A pre-confirmation is roughly a fast early signal that your transaction is in line to be included, before it is fully locked in.

Full transaction finality is targeted at sub-second speeds. Finality is the point where a transaction can no longer be reversed, so getting there in under a second matters a great deal for traders.

To hit those numbers, BNB Chain plans to use an enhanced version of Parlia, the consensus mechanism its existing chain already relies on.

Overall throughput is aimed to exceed 100,000 transactions per second. The research findings attribute that goal to a combination of consensus optimization and parallel execution, which lets the network process multiple transactions at once instead of in a single file line.

Beyond raw speed, the plans include several features that are expected to improve usability:

Native account abstraction, which allows wallets to behave more like programmable accounts.

A shielded privacy pool, intended to give users more transaction privacy.

Improved transaction capacity, rounding out the performance push.

Where the project stands today

NewL1 is not live for the public yet. It currently runs as an internal, multi-validator development network, which the research describes as operational as of October 2026.

A public testnet is anticipated by late 2026. A full mainnet rollout is projected for early 2027.

The project did not appear out of nowhere. BNB Chain first highlighted NewL1 in its H2 2026 technical roadmap, published on July 8, 2026.

That roadmap argued that on-chain infrastructure needs to compete with centralized exchanges on performance. It singled out users who depend on autonomous trading and AI applications as a key audience.

The October documentation update fills in that earlier sketch with specific technical goals.

Why a second chain, and why BNB stays central

BNB Chain is framing NewL1 as an addition to its ecosystem rather than a replacement. The stated goal is to extend existing infrastructure without disrupting the frameworks already in place, including BSC.

CZ affirmed that BNB is set to be the single asset for gas, staking and governance across both BSC and the new layer-1.

What this means

For traders, the pitch is simple. If NewL1 hits its targets, on-chain trading could start to feel closer to the responsiveness of a centralized venue, while keeping assets in self-custody.

A 20 ms pre-confirmation and sub-second finality would narrow one of the biggest gaps between decentralized and centralized trading.

The AI angle is also worth watching. Autonomous agents that trade or transact on their own need fast, predictable settlement, and the roadmap explicitly names them as a target user base.

There are risks. Performance targets on an internal devnet do not always survive contact with a public network, adversarial users and unpredictable demand.

The milestones to watch are straightforward. First, whether the public testnet arrives in late 2026 as anticipated. Second, whether measured performance on that testnet comes close to the 200 ms blocks and sub-second finality promised in the docs.

Then comes the early 2027 mainnet window.

Disclosure: This article was edited by Estefano Gomez. For more information on how we create and review content, see our Editorial Policy.