The people shaping Starknet’s Cairo-native DeFi ecosystem

Starknet’s DeFi TVL sits at $35.51 million, a figure that looks small next to Base’s $5.87 billion but hides a distinct technical bet. The network runs on the Cairo virtual machine, which executes provable programs to generate STARK proofs for Ethereum verification. This architecture allows for transaction fees as low as $0.002 and confirmations under two seconds, metrics that outperform most EVM-based L2s in raw efficiency. The ecosystem is built by developers who write directly in Cairo, a Rust-like language with ownership semantics and a sound type system. They use the Scarb package manager and Starknet Foundry for testing, tools that have matured significantly since the shift from Cairo Zero. The result is a stack where every account is a smart contract, enabling native account abstraction without the ERC-4337 plumbing required on Ethereum mainnet.

Moody and the Ekubo protocol

Moody Salem, a core contributor to Uniswap v3 and architect of v4, founded Ekubo as a product-first DEX. He rejected venture capital funding to avoid developmental shortcuts, bootstrapping the protocol entirely from its inception. Ekubo combines ultra-concentrated liquidity with a "till-based" execution model to minimize redundant token transfers and gas consumption. The protocol uses permissionless extensions, allowing developers to build DCA tools, launchpads, or derivatives directly on top of the core liquidity layer. Unlike Uniswap, where the UNI token offers no direct revenue sharing, Ekubo’s tokenomics link protocol success directly to holder value. Fees flow to the DAO through a RevenueBuybacks contract, which triggers TWAMM orders to sell accumulated fees for EKUBO gradually over a configured window. This mechanism prevents market-moving trades and ensures economic value accrues to stakers rather than just the treasury.

The governance structure is deliberately narrow, with the EVM V3 Core contracts remaining ownerless and immutable. Ekubo, Inc., the Delaware corporation behind the initial build, holds one-third of the total supply and committed to never selling those tokens. In October 2026, the company completed 33 direct ownership transfers of peripheral contracts to designated DAO proxies across 11 EVM chains. However, six contracts on BNB Smart Chain and Monad remain company-controlled, a residual risk that the board accepted without creating a reachability-proof proposal. Production cross-chain message delivery remains unproven, meaning fee withdrawal functions could be unreachable if a messaging route fails. This gap between configured ownership and actual production capability is a clear weakness in the current setup.

Tooling and the zkFi transition

The developer experience has transformed with the release of Cairo 2.6.0, which upgraded Sierra to v1.5.0 and requires Starknet version 0.13.1 or higher. StarkWare is moving toward Cairo-Native execution, replacing VM emulation with native code compiled via MLIR and LLVM. This change removes the overhead of interpreting instructions step-by-step, allowing the sequencer to process transactions significantly faster. LambdaClass conducted extensive replay tests to ensure consistency between native execution and the Cairo VM prover path. For developers needing complex computations, Reilabs introduced Cairo Hints, an extension that allows programs to query external JSON-RPC servers for data like square roots. This moves expensive operations outside the proof system, reducing circuit costs while maintaining verifiable correctness.

The broader vision is a transition to "zkFi," embedding validity proofs at every level of the DeFi stack. Starknet’s architecture supports this by treating proving as a first-class feature rather than an afterthought. Developers can now create highly customizable L3 appchains using the new Cairo Verifier developed with Herodotus. The challenge lies in adoption; the non-EVM environment means established protocols like Aave cannot deploy unchanged. This creates a porting cost that keeps the DeFi stack smaller than Arbitrum or Base. Yet, for teams willing to write Cairo natively, the performance gains are tangible. The question remains whether the technical advantages of STARK proofs will outweigh the liquidity deficit compared to EVM chains.