Original author: Cem | Sovereign
Original translation: TechFlow
Recently, much of the discussion on Crypto Twitter has focused on improving performance in general-purpose environments.
Base is working towards the “Gigagas” goal by leveraging Reth and Ethereum’s blob upgrade.
Solana is getting closer to reaching its vision of 1 million TPS (transactions per second) through Firedancer and amazing optimizations of its C-based network stack.
MegaEth completely removes the gas limit with the help of a highly optimized sequencer.
As a crypto enthusiast, I have nothing to complain about. I paid huge fees on Ethereum mainnet during the DeFi summer and the 2021 bull run, and now I can enjoy low-cost transactions on Solana. And in the future, I will also enjoy cheaper and faster transactions on all these platforms.
However, since I first entered the industry around 2017, I have been obsessed with making crypto mainstream, and lately one question has been on my mind:
We are rapidly approaching the tipping point of over-optimization.
By the end of 2025, block space will be plentiful and performance will become a commodity. Once near-instant and free transactions become the norm, sheer speed will no longer be the key to standing out. As developers, we need to think differently.
Post-performance era
We call this the “post-performance era” because the performance war has been largely won. Most platforms now enable fast and cheap transactions, so differentiation must come from somewhere else: unique features and experiences.
This is where full stack customization comes in. It’s 2025, transactions are cheap and fast, but most apps still look and feel the same. Meanwhile, the market premium for launching another Ethereum Virtual Machine (EVM) has disappeared. Just look at Unichain, which has failed to attract widespread attention or liquidity:
Meanwhile, the winner of this round — Hyperliquid — took a bold approach. It built its entire stack from scratch, optimizing it for specific use cases. Among the many interesting customizations it introduced, two stand out:
Priority Cancellation and Pending Orders Only
By enforcing the order of trades by type within a block, Hyperliquid protects stale orders from being easily snatched up by high-frequency traders. This reduces harmful order flow and makes market making easier, thereby increasing liquidity for all traders.
Vault-based copy trading
Hyperliquid vaults allow anyone to automatically replicate the vault builders trades. Since the vault logic runs as part of block creation, no external maintainer is required. Hyperliquidity vaults run market making strategies, allowing anyone to provide liquidity and share in the resulting profits and losses.
Combine these unique features with high performance, low latency, and a seamless user experience, and it’s easy to see why Hyperliquid has become the DEX of choice for derivatives.
The results speak for themselves:
The real bottleneck: virtual machines
One of the main reasons for the lack of differentiation in most applications is virtual machines (VMs). Many of our tools are built around VMs (or derivative versions of Ethereum and Solana clients), which in some ways hinder customization.
In recent years, there has even been a trend in the industry to push all Rollups to achieve full EVM-equivalence, so that they can be based on the Ethereum Virtual Machine (EVM) and run natively. This is really cool for us tech geeks, but is this what the market really needs?
I dont think so. In fact, the more you can differentiate on features that really impact customers, the more likely you are to win in the market.
Isn’t interoperability key?
Yes, even among specialized blockchains, cross-chain communication is key. Standardized and shared liquidity and messaging tools remain critical.
That’s why open source messaging libraries like @hyperlane and intent-based bridging frameworks like @RelayProtocol are expected to succeed. But developers should have the freedom to fully customize their applications, in addition to still being able to integrate these interoperability components on custom chains.
Responding to market demands
The market needs finely tuned applications built for a specific purpose and providing a seamless user experience - not just another run-of-the-mill EVM derivative. So build something truly custom and optimized for your use case.
Only then can we create applications that truly bring crypto into the mainstream.