New Stack Targets On-Chain HFT Bottlenecks

Altius Labs' core innovation lies in instruction-level parallelism (ILP) combined with optimistic concurrency control (OCC). This approach contrasts with traditional sequential execution models by allowing multiple transactions to be processed simultaneously. Instead of re-executing entire conflicting transactions, Altius's system, using a technique called Static Single Assignment (SSA), identifies and re-processes only the specific conflicting instructions, aiming to reduce wasted computation and gas fees. The team's background includes deep experience in high-frequency trading from firms like Hudson River Trading. This informs their focus on resolving core blockchain inefficiencies such as state access bottlenecks. To address this, Altius employs a scalable storage solution that partitions state data into shards, enabling concurrent read/write operations, a technique designed to prevent the storage layer from becoming a performance bottleneck as network activity grows. This modular execution layer is designed to be VM-agnostic, supporting not just the EVM but also environments like MoveVM and WASM. The goal is to offer a plug-and-play performance upgrade for various L1s and L2s without requiring developers to learn new languages or alter their existing toolchains. The roadmap includes full EVM-compatible integration, followed by support for alternative VMs, and ultimately the launch of a Shared Execution Layer. In the broader competitive landscape, other projects are also targeting high performance with parallel execution. Monad, for instance, pipelines transaction execution to achieve parallelism while maintaining EVM compatibility. Sei Network utilizes a "Twin-Turbo" consensus mechanism and optimistic parallelization to reach high transaction throughput, attracting institutional interest from firms like BlackRock for on-chain experimentation. From an infrastructure perspective, the push for on-chain HFT mirrors advancements in traditional finance. Firms like Jump Crypto are developing solutions such as Firedancer for Solana, which uses a modular architecture and kernel-bypass techniques to maximize hardware efficiency and transaction throughput. This highlights a trend where trading firms are not just using but actively building and optimizing the underlying blockchain infrastructure to gain a competitive edge. For latency-critical systems, the deployment architecture is a key consideration. While cloud platforms offer scalability, on-premises and co-located servers provide greater control over hardware and more predictable network latency, which is crucial for HFT. Modular systems like Altius could potentially be deployed in a hybrid model, where critical, latency-sensitive components are run on dedicated hardware, while other parts of the stack leverage the flexibility of the cloud. The involvement of major financial institutions is rapidly maturing the digital asset infrastructure. Citadel Securities has made strategic investments in Web3 infrastructure, including Ripple and the institutional-focused EDX Markets, and plans to provide liquidity to major exchanges. Similarly, Morgan Stanley is building out its digital asset capabilities, including plans for its own crypto platform and custody solutions, signaling a long-term strategic integration of digital assets into their core services. This institutional push is creating demand for robust, high-performance on-chain infrastructure that can handle sophisticated trading strategies. Startups founded by former quants from firms like Coinbase, Jump, and Citadel are also emerging to tackle challenges like MEV leakage and toxic order flow, further underscoring the convergence of traditional HFT principles with decentralized finance.