From on-chain integration to on-chain rewriting, DeFi is starting to tinker with the underlying infrastructure again.
During the DeFi Summer of 2021, everyone was focused on token issuance, mining, andmicro-innovations; in 2023, DeFi is reconstructing again. But this time, it's not about module integration or gameplay innovation, but rather "reversely" layering from the bottom up.
You'll notice more and more projects are not building on the previous generation's wheels, but questioning:
"Is this wheel designed incorrectly?"
Thus, on-chain trading is beginning to take two paths:
Either do everything yourself, build the chain, write the matching, and handle wallet interactions
Or only write the bottom-layer components, modularize everything, and let others combine them into a system
Today, we'll discuss this ongoing bottom-layer reconstruction battle. Not analyzing projects, but looking at what problems these projects are solving and why we should care about this trend.
Problem One: Why is on-chain trading still not done well?
On-chain trading, starting with the AMM (Uniswap) revolution, once broke through market-making barriers but also shattered efficiency.
If you want depth, you lose efficiency; if you want efficiency, matching must return to centralization.
In recent years, on-chain trading has tried to upgrade from AMM to "on-chain CEX", resulting in either L2 (cheap gas but no users) or creating a chain (created a chain but no one adopts), ultimately discovering that the problem is not TPS, but:
Matching and settlement are not decoupled
Liquidity is fragmented between chains and DEXs
Cross-chain trading experience is poor, with cumbersome wallet interactions
So now, the direction is not "a better DEX", but directly rebuilding the trading system's foundation.
Hyperliquid: On-chain trading system might not need layering at all
Hyperliquid's approach is:No L1/L2 separation, no matching/settlement separation, simply create a native high-performance chain and directly write matching and trading modules into the chain logic.
Benefits include:
Matching processed on-chain, trades verifiable
No need for Sequencer, no external settlement nodes
All assets and liquidity aggregated in a unified account system
Simply put:
"Not hanging a DEX on the chain, but making the chain itself the exchange."
This approach is somewhat like Solana, but without VM, directly native and customized for trading. The cost is high coupling and poor scalability, but the experience is truly smooth.
And Orderly is more like a multi-chain form of Hyperliquid, following a "rural surrounding urban" strategy—not dominating a single chain, but using modularity and multi-chain layout to enable more chains and projects to use "native exchange" level performance and liquidity.
(Translation continues in the same manner for the rest of the text)This is completely different from the early DeFi's "attracting liquidity through subsidies" logic, more like a traditional financial HFT strategy fund - just becoming an on-chain version with a lower threshold.
DeFi is not becoming simpler, but more "systematically engineered"
Today's DeFi is far from the situation of "modifying contracts and mining" in 2020.
Either you do like Hyperliquid with full-chain self-research, pursuing extreme performance;
Or like Ethena, combining on-chain tools into "real financial scenarios";
Or like Orderly, building standard components, allowing others to quickly assemble products.
These three approaches are not right or wrong, but are each making up for DeFi's "engineering structure shortcomings".
Future hit products may not necessarily create their own chain or issue their own token, but they will definitely leverage these structures.
If you are still concerned about "which coin can rise", you might have already missed this round's deepest narrative.
The protagonist of this round is not the coin, but the structure itself.
