On June 25, the Base network goes through a change that the average user will never feel as an event, yet one that could quietly reshape how digital assets are created on one of the busiest Layer 2 networks in the Ethereum ecosystem.
The upgrade is called Beryl, it already runs on the Base Sepolia testnet, and it rests on three substantial shifts. It brings a token standard that lives inside the network software itself, it shortens the time a user waits to withdraw funds back to Ethereum, and it readies the infrastructure for a higher volume of transactions.
Taken one by one, the three changes sound dry and purely technical. Put together, they make Coinbase’s direction for the network it built fairly easy to read.
What Base is and why anything on it matters
Base is a Layer 2 network developed by Coinbase and launched on mainnet in August 2023. The idea behind a Layer 2 is simple enough to follow even for someone who has never opened a crypto wallet. Ethereum, the base chain, is secure and decentralized, but it turns expensive and slow when millions of people use it at the same time.
A Layer 2 carries the load. It runs transactions off the main chain, bundles them, and then sends only a cryptographic summary back to Ethereum. The user pays a fraction of the usual cost, while security stays anchored in the parent chain.
Base belongs to the family of networks known as optimistic rollups, which start from the assumption that transactions are valid until proven otherwise. It was originally built on the OP Stack, the open source infrastructure developed together with Optimism, and it has no token of its own.
Fees are paid in ETH, a deliberate choice Coinbase made from the start to keep the experience simple and to discourage the speculation that usually surrounds a new coin. For an audience used to the dozens of tokens that appear and vanish within a single season, that restraint has become part of the project’s identity.
The fact that Base now ships upgrades a few weeks apart says something about its technical maturity. A fragile network cannot afford to swap out core components without breaking something along the way.
The full breakdown of this upgrade was covered in depth by Cryptology.ro, the Romanian crypto news and analysis publication, where Mihai Popa keeps a close watch on how Layer 2 networks evolve. For readers who want to see how these technical shifts turn into concrete opportunities, the newsroom regularly publishes guides on the most profitable cryptocurrencies and on where the market is heading.
Beryl arrives just four weeks after Azul
Beryl makes more sense if we glance back at the upgrade that came before it. Azul, released four weeks earlier, moved Base onto a fresh foundation called the Base Stack. It was a rebuild of the protocol with two aims. It simplified the code, which makes the network easier to audit and harder to break, and it introduced a system that relies on more than one proof, known as multiproof, which confirms withdrawals through a combination of methods rather than leaning on a single one.
Azul was, for the most part, the groundwork. Beryl is the floor raised on top of it. The Base team openly admits that the four week pace between the two upgrades would have been all but impossible before the migration to the Base Stack. In a landscape where dozens of Layer 2 networks fight over the same liquidity and the same developers, the speed at which you can adjust the protocol without incidents has become a selling point in its own right.
B20, the standard that lives at the heart of the network
The headline feature is B20, the native token standard of the Base network. The name deliberately echoes ERC-20, the standard that governs most tokens on Ethereum, and the resemblance is anything but accidental. B20 was designed as an alternative that keeps everything that works about ERC-20 and changes only where the code lives.
A precompile, not an ordinary contract
This is where the whole story sits. A classic token on Ethereum or Base is a smart contract, a program someone uploads to the chain that runs inside the Ethereum virtual machine. Every time someone sends that token, the network executes the contract’s instructions step by step, exactly as it would any other piece of code hosted on chain.
B20 works differently. Its logic does not sit in a contract layered on top of the network but is built directly into the node software, written in Rust and executed natively. In technical terms, B20 is a precompile, a kind of system function the network knows by design and can run far more efficiently than an arbitrary program.
The closest comparison from everyday computing is the difference between an app a user installs and a component included in the operating system. Both do the job, but the second is faster, better integrated, and harder to disrupt from the outside.
Compatible with ERC-20 down to the last detail
A natural question follows. If B20 works differently under the hood, what happens to the wallets, exchanges, and applications built for ERC-20 over the years. The answer is the clever part of the design. B20 fully implements the ERC-20 specification, which means a token created under it behaves, seen from the outside, exactly like any ERC-20 token.
For a developer, that means there is nothing to rewrite. A B20 shows up in MetaMask, gets recognized by block explorers, can be listed on exchanges, and interacts with existing DeFi protocols without any special integration. The technology behind it has changed radically, while the surface the rest of the world touches stays untouched.
That continuity is exactly what gives B20 a real shot at adoption rather than condemning it to the status of a technical curiosity. Crypto history is full of standards that looked superior on paper and failed because they asked the ecosystem to start over.
The issuer toolkit and the compliance controls
B20 is not only a faster way to move tokens. It ships with what Base calls the Issuer Toolkit, a set of tools built for companies that want to issue regulated assets, from stablecoins to tokenized real world assets. The team noticed that issuers kept rebuilding the same compliance features from scratch every time, which slowed their launches and left room for costly mistakes.
The toolkit hands over the controls a serious issuer needs from the very beginning. It can mint and burn tokens, set a supply cap, pause transfers, and enforce rules that allow or block specific addresses. A more delicate function lets it burn funds held by a blocked address, the mechanism known in regulatory language as freeze and seize, which authorities often demand when investigating assets obtained illegally.
Base has been transparent on the point, stating that freezing decisions rest with issuers alone and not with the network, and the entire toolkit went through an audit carried out together with the security firm Spearbit.
The same reading runs through the analysis written by Mihai Popa, an analyst and columnist at Cryptology.ro, who argues that the real stake in Beryl is not the headline speed numbers but Coinbase’s ambition to turn Base into a serious venue for issuing regulated assets.
Withdrawals drop from seven days to five
The second major change touches a sore point for anyone who has tried to move money off a Layer 2 back to Ethereum. The process is not instant, and the reason lies in the nature of an optimistic rollup. A withdrawal has two stages, initiation and finalization, separated by a delay the protocol imposes.
That delay is not bureaucracy, it is a security guarantee. Because the network assumes transactions are valid until proven otherwise, it needs a window in which an honest participant can challenge a fraudulent withdrawal before the money leaves.
The classic seven day window dates back to when Base relied on a proof system based on challenges. Azul changed the equation through multiproof. Finalization now requires a positive proof, of the TEE or zero-knowledge type, so the remaining delay no longer serves to wait for a challenge but plays the narrower role of giving the network time to detect and disable a faulty proof mechanism.
Since the window’s purpose has shrunk, its length can drop without putting security at risk. Base lowers the bar to five days and says it will keep going gradually as its monitoring systems mature. For the user, the benefit is plain, two fewer days of waiting on most withdrawals.
Reth V2 and more room for transactions
The third component makes less noise but carries just as much weight over time. Reth is the execution client Base has used exclusively since Azul, software built by the Paradigm team and praised for its speed. Reth V2 targets two of the most pressing constraints on a high traffic network, the disk space a node needs and the speed at which it computes state roots.
The new storage version visibly reduces the space taken up across all node types and gives the network room to grow. The direct consequence is that Base can raise its per block gas limits without overloading the sequencer, which means more space in every block for developers’ transactions.
What users actually need to do
For the vast majority of people who use Base, the answer is reassuringly short. Nothing. The only change they will feel is the shorter withdrawal time back to Ethereum. Node operators are the ones who must update their software to the latest versions before the June 25 activation, or risk falling behind the network, and for them there is good news too, a lower resource footprint thanks to the storage optimizations in Reth V2.
Developers can now launch their own tokens through B20, choosing between the general purpose Asset variant and the Stablecoin variant tailored for fiat backed assets.
Cobalt and where Base is heading
Beryl is not a destination but a stage in a series of upgrades the Base team now controls end to end. The next step is called Cobalt and is scheduled for September. From what has been announced, it will bring native account abstraction, fee payments made directly in B20 tokens, lower transfer costs and higher throughput, plus the merging of the consensus client and the execution client into a single node binary.
The wider picture reveals a clear ambition. Coinbase does not want Base to remain a cheap network for transactions but a first rate platform for issuing assets, able to host regulated stablecoins and real world assets as easily as it hosts experimental tokens today. B20 is the centerpiece of that strategy, and June 25 stays the date the transformation began.
The original reporting and the full analysis by Mihai Popa are available on Cryptology.ro, here.
FAQ
What is the Beryl upgrade on Base?
Beryl is the second major upgrade of the Base network, set to activate on mainnet on June 25, 2026. It brings three substantial changes, the B20 native token standard, a reduction of the Ethereum withdrawal time from seven days to five, and the scalability gains in Reth V2. The upgrade is already live on the Base Sepolia testnet.
What is the B20 token standard and how does it differ from ERC-20?
B20 is the native token standard of the Base network. It fully implements the ERC-20 specification, so tokens built with it work in wallets, exchanges, and DeFi applications without special integration. The key difference is where the code lives. Instead of being a smart contract layered on top of the network, B20 logic runs natively inside the node software, written in Rust and executed as a precompile, which makes it faster and more efficient.
Why does the withdrawal time fall from seven days to five?
The seven day window came from the old challenge based proof system, where the long delay left time to dispute a fraudulent withdrawal. After the multiproof system arrived in the Azul upgrade, finalization requires a positive proof, and the remaining delay only serves to detect and disable a faulty proof mechanism. Because the window’s role has narrowed, Base can safely lower it to five days, and it will continue gradually.
Do users need to do anything for the upgrade?
No. Regular users do not need to take any action. The only change they will notice is the shorter withdrawal time through the canonical bridge to Ethereum. Node operators are the ones who must update their software to the latest versions before the June 25 activation.
Does Base have its own token now that B20 exists?
No. B20 is a standard that issuers use to create tokens, not a token of the Base network itself. Base continues to use ETH for fees and has not launched a native token.
What comes after Beryl?
The next upgrade is called Cobalt and is scheduled for September. It is expected to bring native account abstraction, B20 improvements including fee payments in B20 tokens, and the merging of the consensus and execution clients into a single node binary.
