SVM Expansion: The Landscape Beyond Solana
Not too long ago, the “Modular vs Monolithic” debate dominated conversations throughout crypto. Rollups and Solana seemed completely antithetical, representing polar opposite ends of the scalability spectrum. Today, the conversation has greatly shifted: with the growing number of modular blockchains being developed, more teams are looking to harness the power of the SVM for building highly performant networks.
Layer 2s aim to scale the performance capabilities of Layer 1 base layer blockchains by executing transactions and posting their data back to the base layer for settlement. Naturally, Layer 2s first proliferated on Ethereum thanks to its rollup-centric roadmap and being the first-ever runtime environment for smart contracts. Solana’s resilient rise, however, to one of the largest activity hubs in crypto today has captured the attention of many teams building in the space as a viable alternative.
SVM: Parallel Execution
Solana took a fundamentally different approach to scalability since its inception, prioritizing high performance with minimal latency and costs to users. This approach ultimately optimizes software to best take advantage of performant hardware. Under the hood, a key component in enabling Solana’s performance is the Solana Virtual Machine (SVM), the network’s core execution environment.
While Bitcoin and the EVM adopt a sequential transaction processing mechanism, the SVM was built around processing and executing transactions in parallel, providing Solana with its renowned high throughput (~800 TPS) and fast execution (400 ms). This parallelization is enabled by Sealevel, Solana’s parallel smart contracts runtime environment. Transactions on Solana specify which accounts on the chain they need to access ahead of time, enabling Sealevel to process multiple transactions simultaneously assuming they do not overwrite each other’s actions and create conflicting state.
As Prasad Mahadik highlights in “Solana The Modular”, Anza’s new SVM API decouples the SVM from Solana’s validator runtime environment, marking a critical milestone for the emergence of SVM L2s. Developers can now harness the SVM for building their own high performance networks, with modular flexibility for settlement, data availability (DA), or consensus environments.
The SVM L2 Landscape Today
While still early in its early stages, SVM expansion has taken form in several directions:
- SVM Execution on Ethereum: L2s that settle to Ethereum but leverage the SVM for execution.
- App-Specific Customization: Applications that launch SVM L2s configured to their specific needs.
- Extendable SVM Environments: Custom environments which extend the capabilities of the SVM beyond rollups.
SVM Execution on Ethereum
Performance is often the biggest pain point for EVM L2s today, whether it’s dealing with delays in transaction execution or maintaining low fees while supporting increases in network activity and demand for blockspace. This has spurred a great deal of interest in leveraging the SVM for execution. These SVM L2s combine the existing liquidity & economic security of Ethereum — as well its developer base & tooling — with the high-performance capabilities of the SVM.
Eclipse
Eclipse is the first blockchain to leverage the SVM outside of Solana.
More specifically, the Eclipse mainnet utilizes the SVM for execution, posts transaction data to Celestia (DA), settles to the Ethereum L1, and uses RISC-Zero for ZK fraud proofs. Tapping into the SVM enables Eclipse to take advantage of Solana’s parallel execution, local fee markets, and state growth management.
In sum, this means Eclipse will deliver faster transaction speeds and higher throughput with lower costs than most EVM rollups today, and could get even more performant as Eclipse Labs plans to adopt the highly anticipated Firedancer implementation of the SVM.
To enable easier onboarding for Solana developers, Eclipse is fully compatible with all the standard Solana tooling, including programming languages (Rust, Python), IDE integrations, testing environments, and wallet connectivity. In addition, Solana mainnet programs can be deployed to Eclipse with the Solana CLI. The Eclipse mainnet has been open to developers since late July, and is currently scheduled to open to general use in October. Hyperlane support for Eclipse is already live, available in usenexus.org.
Atlas
Ellipsis Labs recently introduced Atlas, an L2 which will settle to Ethereum and use a custom implementation of the SVM for execution. At a high-level, user transactions are sent to the Atlas RPC, which forwards the request to the Sequencer Node. The Sequencer executes the transaction and broadcasts the next block and its associated state changes to a Replay Node, which is then responsible for replaying the transactions and validating that their computation matches that of the Sequencer.
Leveraging the SVM enables several key breakthroughs for Atlas. For starters, existing Solana programs can easily deployed “out-of-the-box,” allowing Solana developers and applications to easily onboard to Atlas and make the experience as similar to deploying on Solana as possible.
Atlas aims to support 50 ms block times with rapid state merklization, a modular RPC architecture, and a specialized execution environment for executing transactions in a way that makes managing state changes secure and predictable. Atlas is currently running a private testnet, and developers interested in trying it out can fill out this form.
Soon (Solana Optimistic Network)
SOON aims to transform the SVM L2 ecosystem with two key innovations: the SOON mainnet, a versatile Layer 2 , and the SOON stack, a modular rollup framework designed to scale SVM performance.
The SOON mainnet is an L2 that settles on Ethereum, offering a customizable Data Availability (DA) layer that can switch between Celestia, EigenDA, or Avail. It leverages the OP Stack for proof generation and native bridging. The team has engineered a Decoupled SVM framework, where the TPU (Transaction Processing Unit) is detached from Solana’s consensus and controlled by the rollup node. This innovation removes the need for vote transactions, reducing unnecessary DA consumption and improving efficiency for L2s.
A key innovation is SOON’s native ERC-20 <> SPL token bridge, which will allow transfers of ERC-20 tokens directly to the SOON network. In particular, this bridge will enable users to move assets between Ethereum and Solana easily.
In parallel, the SOON Stack is an open-source rollup framework that merges the OP Stack with the decoupled SVM to power high-performance SVM-based L2s. Rollups created with the SOON Stack, called SOON Chains, can deploy to any supported base layer, providing unmatched flexibility. Looking ahead, SOON plans to integrate Firedancer once it’s live on the mainnet, boosting the performance capabilities of both the SOON mainnet and SOON Stack.
Bitcoin L2s Have Entered The Chat
Beyond the EVM, there has also been a great deal of interest around tapping into the SVM from the Bitcoin L2 ecosystem. Hyperlane is already live on EVM L2s on Bitcoin like BOB and Merlin. Similar to EVM L2s, Bitcoin L2s are looking to get “the best of both worlds” by harnessing the deep liquidity and economic security of the Bitcoin base layer, while benefiting from performance capabilities enabled by the SVM.
While this landscape is still in early stages, there are already some noteworthy developments. Yona is an L2 that settles to Bitcoin while using the SVM for execution, allowing users to operate within the Bitcoin realm and avoid the need for wrapped assets. Molecule is launching a multi-faceted product suite, which will include an SVM-based zk rollup, a decentralized sequencer network for SVM rollups, and a settlement-agnostic SVM SDK for appchain developers.
Extending Solana Through Customization
A growing number of Solana application developers have also been exploring building their own SVM L2s. The primary motives for doing so are distinct from that of L2s from external ecosystems — who benefit from significant performance improvements with the SVM as opposed to their native VMs.
Solana application developers are instead focused on launching L2s for economic and app-specific cases: custom block space that doesn’t compete for inclusion with other events on the base layer, implementing custom value accrual mechanisms, and other configurations for achieving next-level performance without overwhelming the Solana L1. Building on a custom L2 also enables teams to ship infrastructure updates much more efficiently as opposed to waiting for Solana, which has become too big of a ship to turn as quickly.
Onchain CLOB DEX: Bullet and Zeta X
Zeta Markets, a Solana-native perpetuals DEX, recently announced Bullet, a “real-time L2” on Solana. Unlike the projects mentioned earlier, Bullet won’t be using the SVM, and will instead leverage its own zkVM for execution while posting data availability and settling to the Solana L1. Bullet’s flagship application will be Zeta X , or ZX, an optimistic rollup offering a fully onchain CLOB DEX. Traders on ZX submit transactions to a sequencer node responsible for ordering transactions, matching trades, and producing blocks. Once transactions are ordered and executed, the sequencer will then publish ZX blobs to the Solana DA smart contract for settlement.
Having been around since 2021 and achieving over $11b in lifetime volume on Solana, Zeta Markets envisions ZX as a solution for providing the optimal ‘CEX-like’ trading experience while retaining the core advantages of onchain trading. ZX is being built in close collaboration with the Sovereign Labs team, and its mainnet launch is currently scheduled for Q1 2025.
Ephemeral Rollups: MagicBlock
MagicBlock is introducing a novel approach to blockchain scalability through ephemeral rollups. This resembles a rollup-on-demand model but operating at the account level: the parallelization of the SVM allows only specific accounts (state) to move into the ER session, while retaining its other transactions on the Solana L1. This is made possible by enabling users to lock arbitrary accounts on Solana and launch a dedicated runtime operating only on those accounts. This model enables a wide variety of use cases, from real-time fully onchain games to CLOB that keep liquidity on Solana and run the matching engine on the Ephemeral Rollup.
Developers continue to deploy programs and assets on the Solana L1, maintaining full composability with existing programs. Meanwhile, transactions that are executed offchain benefit from validators co-location (for real-time latency) and runtime customizations. This approach intends to accommodate all use cases that require real-time latency (games, finance etc) without liquidity or composability fragmentations for developers and users.
Shared Sequencing: Hypergrid & Sonic
Hypergrid is an SVM rollup framework which introduces semi-autonomous environments known as Grids. Grids are app-specific networks designed for high-demand use cases such as gaming, while leveraging the Solana L1 for consensus and finality. Each Grid has its own zk-Coprocessor, SVM runtime, gas engine, and merkle tree generator. The HyperGrid Shared Sequencer Network (HSSN) contains a network of multiple validators that manage settlement of zk proofs for Grids. Grids can service different categories of users with different needs and preferences, while the HSSN ultimately serves as the consensus layer, communication hub, and state management cluster for all Grids.
Sonic is the first atomic SVM chain built on top of the Hypergrid framework, offering a highly performant, low-cost onchain gaming experience among all gaming L1s. Transactions can be executed on Sonic without the need for developers to redeploy their Solana programs and accounts. With the HyperGrid interpreter, dApps from EVM chains can also be deployed to Solana. Sonic launched its Odyssey testnet this summer, and recently announced their $12m Series A raise, so its likely their mainnet will launch sometime in Q4 ’24 or Q1 ‘25.
Decentralized AI: Grass Network
Grass is building a decentralized network for scraping public web data and transforming it into structured datasets used for training AI models. By enabling anyone to run a node on an everyday device such as a laptop or mobile phone, Grass needs to handle tens of millions of web scraping requests per minute, a scale which no L1 can currently support.
A key component of ensuring data collected is reliable is assigning verified metadata to each dataset, detailing its source, collection time, and processing methods. To store all of this data in an efficient and scalable manner, Grass is launching the first L2 data rollup. The Grass L2 will function as a zk rollup, batching metadata into proofs for settlement and storage on the Solana L1. Though no dates have been confirmed yet, it’s believed that Grass will launch by the end of 2024 or into early 2025.
Extending The SVM Beyond Rollups
Though rollups often dominate discussions around scalability, the scope of SVM expansion extends beyond them. For instance, Termina is an “SVM-as-a-Service” platform that allows anyone to easily deploy custom Solana environments as SVM rollups, Batchers, or SPEs. Though it sounds like a lot, the distinctions are relatively simple to understand.
- Batchers — Batchers are networks which aggregate transactions offchain but execute them on the Solana L1. Although this creates a delay in execution, user’s funds always remain on the base layer. Notable teams working on Batchers include Code, a peer-to-peer payments network, and Cube, a hybrid onchain — offchain exchange.
- SPEs — As defined by the Solana Foundation, Solana Permissioned Environments, or SPEs, are “customizable Solana Virtual Machine private instances”. This model targets enterprises and institutional clients who prefer to retain full control over critical network infrastructure while running Solana code. In particular, SPEs allow full control over validator sets, chain functionality, and gas token implementations. Pyth Network’s Pythnet is the first live SPE, which uses its own custom validator set and PGAS token, while harnessing the SVM’s parallel processing power to connect to various external data sources at scale. Other notable examples include Iron, who is developing an onchain bank, and Rimark, who is developing an SPE for asset tokenization in fintech applications. Most recently, Sphere Labs unveiled SphereNet, which will be a permissioned environment for coordinating activity between regulated financial institutions across the globe.
Additionally, while the SVM’s execution capabilities have been the center of attention, there has also been interest brewing around extending other parts of the Solana base layer including its consensus layer. For instance, Rome Protocol is developing an extensive suite of products and services for L2s powered by Solana. This will include a Shared Sequencer Network, custom EVM implementation, Rome L2, and more. Rome cites fast transaction confirmation times, high throughput, and immutable transaction ordering guarantees as some of the most critical reasons for using Solana.
SVM Expansion: Accelerated By Hyperlane
2+ years ago, Hyperlane’s core contributors made a bet on the expansion of modular blockchains. The key insight was that by separating the core functionalities of blockchains (consensus, execution, settlement, data availability), developers were no longer restricted to the capabilities of any single base layer. Instead, they could pull different components from different architectures to best fit their case-specific needs. Fast forward to today, and the modular blockchain landscape is alive and thriving with many new L2s and app-chains live or in production. Now, Hyperlane sees a similar expansion presenting itself with the rise of SVM L2s.
SVM expansion is well underway, but not all SVM L2s are created equal. While some may settle to Solana, others settle to Ethereum or Bitcoin instead. Perhaps in the future there will be SVM L2s that settle to somewhere completely new, or SVM appchains, or SVM sovereign rollups.
The big question emerging here is how can all of these networks communicate with one another across these different environments?
Hyperlane completes the modular blockchain stack with permissionless interoperability. Built to be environment-agnostic as an open source framework, Hyperlane offers the widest possible range of chains for new chain deployers and asset issuers to connect to from day one, without imposing any vendor lock-in over core infrastructure components.
Overall, there are several key reasons why Hyperlane is the ultimate interoperability solution for teams actively building SVM L2s:
Permissionless Deployments = Faster Growth
With Hyperlane’s expansion into the SVM, anyone can now permissionlessly deploy Hyperlane on any new SVM L2, enabling connectivity between L2s, Solana, and any other ecosystem. And once connected, L2 & app devs have the freedom to create any token bridge or customize the security method used.
Deploying Hyperlane to a new environment is an entirely permissionless process — you can add a new chain to the network by deploying out-of-the-box. Hyperlane provides the necessary developer tooling for chain deployers and asset issuers to set up their own token bridges (called Warp Routes) including for SVM, as well as prebuilt Warp Route UI templates.
Furthermore, as the SVM landscape continues to grow, it’ll need to draw developers & liquidity from multiple different environments, emphasizing the need for seamless cross-VM interoperability. An SVM L2 that settles to Ethereum needs to be able to easily communicate with Solana or an SVM L2 that settles to Solana. As the most extensible interop protocol today, Hyperlane is available across more VMs than any other existing solution.
By significantly reducing bottlenecks around interchain infrastructure deployment, teams can dedicate their best resources and efforts towards their domains of expertise. And for those that don’t want to handle any interop work, there’s optional hands-on support available from Hyperlane’s core contributors.
Expand Anywhere with Warp Tokens
With Warp Tokens, asset issuers can take part in SVM expansion across the permissionless chains and VMs available with Hyperlane. Warp Tokens are tokens — fungible or non-fungible — that can be easily sent and transacted across Hyperlane-connected chains.
With Hyperlane, chain deployers can set up synthetic token routes to create new warp tokens on destination chains that represent tokens from the origin chain. When synthetic warp tokens are minted on Solana, they can be minted as standard SPL assets or Token2022 assets, allowing builders on Hyperlane to configure additional logic and functionality for their assets if they wish to do so.
Future-Proof Security
As there are many new SVM L2s and L2 frameworks on the horizon, there will be a dire need for secure and seamless interoperability between them all. With Hyperlane’s modular security architecture, teams can configure their own security preferences as they wish by selecting from Hyperlane’s extensive library of Interchain Security Modules (ISMs).
What’s powerful about this structure is that it aims to be future proof: teams can always specify which verification mechanism they would like to use, therefore as new interoperability verification mechanisms become available, they can opt to use those.
For instance, there has been previous interest and research around Diet Clients on the SVM, which would enable lightweight verification of transactions and network integrity by relying on fraud proofs. This mirrors how light clients support verification on IBC today. If this idea becomes a reality in the future, anyone can add a Diet Client as a Hyperlane ISM in the future.
Beyond Diet Clients you could use Hyperlane ISMs to set up all kinds of custom security: EVM rollup bridges or even, your own app-specific validators & economic security, and rate limiting. Again, white glove support from the core dev teams is available directly or on #developers in Discord to help you implement Hyperlane to meet your security needs.
SVM Expansion in Action
Just last week, Hyperlane set foot into the world of SVM expansion with Eclipse, the first SVM L2 outside of Solana. As an L2 leveraging both SVM execution and EVM settlement, Eclipse needed to connect assets which are native to both environments.
In a collaborative effort between Hyperlane and Eclipse contributors, Hyperlane now powers the primary bridge to and from Eclipse, allowing seamless asset transfers between the EVM and SVM. Users can now bridge into Eclipse via Ethereum, but can withdraw out of Eclipse to Solana instead. This is just one of the many examples of what Hyperlane enables for teams building SVM L2s today. Read more about the Eclipse integration here.
Hyperlane offers developers the fastest and easiest way to take part in SVM expansion today. Permissionless deployments and customizability that Hyperlane offers makes it the most scalable solution for the incoming demand for interoperability among SVM L2s, supported by the most extensive connectivity across chains and VMs currently available. Moving forward, new SVM chains can easily add interoperability and fast token bridging with Hyperlane, joining over 60 different chains across 6 different VMs in the Hyperlane ecosystem today.
SVM Expansion.
Shoutout to Prasad Mahadik at Delphi Digital for his work on Solana The Modular. And a big thanks to Garrett Harper from Squads, Ren from Eclipse, Rahul from SOON, Andrea from MagicBlock, Jarry Xiao from Ellipsis Labs, Joe Caulfield from Anza, and Mason Nystrom from Pantera Capital for feedback and support.
More about Hyperlane
Hyperlane is the open interoperability framework. It empowers developers to connect anywhere onchain and build applications that can easily and securely communicate between multiple blockchains. Importantly, Hyperlane is fully open-source and always permissionless to build with.
