The conversation mainly revolved around Solayer's technical solutions. Solayer has the strategic goal of building a high-performance blockchain with millions of TPS, and is driven by hardware acceleration and protocol reconstruction: At the technical implementation level, Jason introduced in detail the optimization path of Solana's verification nodes, including the architectural innovation of InfiniSVM - through parallel processing enhancement and state compression mechanism, it achieves 4 times the efficiency improvement compared to traditional EVM and Solana SVM; in terms of ecological construction, Jason emphasized the market positioning of Solayer InfiniSVM and future development plans.
Personal experience of Jason, a technician at Solayer
Jason: My name is Jason Li. I have more than three years of experience in the blockchain industry. I am currently the technical director of Solayer, mainly responsible for the development of the public chain and the construction of the initial Restaking Protocol.
When I was in high school, I conducted a comparative study of cryptocurrency and traditional asset price trend models in Singapore. After graduating from the University of California, Berkeley, I worked on AI training optimization research and developed a large model training scheduling algorithm called Tik-Tok Scheduling Algorithm (the result was cited by the Google team), which aims to improve CPU utilization during large model training and achieve additional calculations in the IO stage. Subsequently, I conducted research on the Multiparty Computation algorithm, and a patent has been applied for in the United States. The full name of the patent is Efficient Hardened Derivation in Multiparty Computation Setting, which is the basis of the NPC Vault technical architecture.
The opportunity and goal of Solayer’s creation: What problems does the Solana ecosystem need to solve?
Jason: The starting point of Solayer's creation is an in-depth exploration of the scalability and shared security of the Solana ecosystem. Solana is naturally endowed with significant transaction speed and user experience advantages, which makes it significantly ahead of other public chains in the current market cycle. Whether it is a high-concurrency transaction scenario such as Memecoin or a Web3 application ecosystem for ordinary users, Solana can well release the network effect brought by its technical characteristics. Based on this judgment, we chose to first focus on improving the yield of basic assets through Resttaking and Validator performance optimization, rather than simply copying the revenue model of projects such as Eigen that focus on security services (such as AVS). At present, by optimizing the transaction packaging efficiency of Solana validator nodes, we have made the SOL staking scheme one of the highest-yielding liquidity staking tokens in the ecosystem. In addition, we are also working with OpenEden to increase the yield of US dollar assets (such as USDC) in a more stable way, rather than pursuing short-term AVS returns.
At the same time, we also clearly see that although Solana itself has good scalability, there is still a lot of room for improvement in the trading experience. Especially in high-concurrency scenarios, such as popular token transactions or Memecoin transactions, users often encounter problems such as transaction delays, additional fees, and even malicious MEV arbitrage. Solayer is actively solving these bottleneck problems through hardware optimization, the introduction of AI performance improvement solutions, and the redesign of network and consensus mechanisms to achieve a truly real-time interactive trading experience. Our ultimate goal is to achieve a performance of one million transactions per second (TPS). At present, through phased efforts, we have achieved about 500,000 TPS in the actual operating environment (explorer.solayer.org), and with the completion of subsequent optimization work, it is expected to reach 1 million TPS.
What is InfiniSVM?
As SVM Layer1, InfiniSVM improves the speed, scalability and security of blockchain systems through hardware acceleration. Unlike traditional blockchain architectures, InfiniSVM's vision is to achieve infinite scalability, ensuring that the network does not sacrifice security or efficiency while processing a large number of concurrent transactions and smart contract operations. One of its design goals is to achieve a million TPS (million transactions per second), greatly improving the throughput of the blockchain network and meeting more complex and high-frequency application requirements.
The launch of InfiniSVM is expected to bring unprecedented breakthroughs to blockchain technology, especially in application scenarios with high throughput and high security requirements.
How to understand InfiniSVM and its future user scenarios?
Jason: Overall, blockchain technology has improved by several orders of magnitude from Bitcoin to Ethereum, and Solana is also much faster than Ethereum. What we are pursuing is whether Solana can be several orders of magnitude faster on its existing basis, rather than just increasing from 7,000 to 10,000, which is not very meaningful to us, so we directly set the goal at 1 million TPS.
Technically, this is undoubtedly a direction for market development, but I cannot predict which specific applications will be implemented. For example, in terms of trading applications, we have seen some very active products on the chain, and their user experience has been greatly improved; and as for whether the meme project will become the best application on the chain, I am not sure, this can only be tested by the market.
I also think that the demand for blockchain is growing, and more and more people are willing to use blockchain and put all kinds of assets on the chain. At the same time, they want to have a user experience similar to Web2, instead of having to worry about network congestion, adjust fees, or call other APIs to estimate transaction parameters every time they operate. Sometimes Solana is still not on the chain 60 seconds after you initiate a transaction, or when you need to do various complicated operations on Ethereum to exchange, the whole process makes users feel confused. These are problems we strive to avoid.
I think this is exactly what InfiniSVM can provide. Although I can't foresee which applications will really explode in the future, I can be sure that the order of magnitude performance improvement will definitely give rise to some very interesting applications.
What is the thinking behind the strategic shift from Restaking to InfiniSVM?
Jason: Initially, we thought that Restaking is essentially an intermediate component to improve scalability. Unlike EigenLayer's strategy, we did not choose to wait for the formation of the trading platform ecosystem before supporting the AVS ecosystem, but directly committed to achieving scalability. This is the core difference between the two of us, because Solana does have this urgent need. Solana often freezes during transactions. For example, my friends and I often encounter problems such as long-term unsuccessful transactions and waiting for 60 seconds to confirm when transferring money. From the user's perspective, these problems must be solved to bring about real qualitative changes.
Solana is a very unique platform to me, it predates Aptos and Sui, both in terms of the VM layer and the interaction layer, and has made many significant innovations. Bugs are inevitable in such a large system, and this is a process of continuous iteration. More importantly, we need to focus on the experience that users really need and love. If you look at the on-chain activity, you will find that there is not only PumpFun, but also many Memecoin Launchpad projects, and even protocols launched directly from Meteora or DLMM. If these protocols are copied to Ethereum, they may not have the same effect, and Ethereum's USDC is constantly migrating to Solana. This shows that from a user perspective, everyone prefers this efficient and smooth experience. If this experience direction proves to be excellent, we will take another big step on this road. This is our idea.
What is unique about InfiniSVM compared to EVM, Solana SVM, and Sonic?
Jason: The advantages of InfiniSVM over EVM are very obvious. The architecture of EVM leads to its lower throughput because it adopts an optimistic approach, such as using an optimistic approach like Optimism, but due to the architectural design, even with this approach, the conflict rate is still very high. In other words, for the same type of transaction, when using EVM on Ethereum and when we use SVM to process, SVM will have a higher throughput due to design reasons. This is not only because of the difference in execution mode from Optimistic to Pessimistic, but also mainly due to the difference in state access - in EVM, tokens are only defined as an interface, while on Solana, tokens exist as first-layer citizens, which directly affects the difference in TPS.
In addition, the user experience of Ethereum is not particularly good. For example, when using a certain login protocol, users need to click approve first, and then click swap; for another example, in the cross-chain bridging process, assuming that you are bridging from Ethereum to Arbitrum, when using DeFi applications on Arbitrum, users need to approve first, then click bridge, and then wait for a while, and then click approve again on Arbitrum. In this protocol, you also need to click deposit once more. In total, you may need to click four times and wait for four transaction cycles to complete the transaction.
Under Solana’s SVM architecture and compared to Sonic SVM, we can allow users to complete all operations with just one signature. This design allows users to interact with the application without feeling the complicated process behind it, and only feel that the whole experience is very smooth.
Compared to the public chain, Sonic mainly focuses on industry-customized applications, mainly focusing on the gaming track. But the needs for games are somewhat different. For example, if I have a game, I can deploy a chain using hypergrid architecture - this is my understanding. Basically, they want to make a customized SVM expansion chain to Solana, so that everyone engaged in games can deploy an independent side chain to control the content on the chain and customize it. This is Sonic's goal. Solayer's goal is different. We want to build a universal chain that allows the entire chain to run extremely fast. We have made a lot of underlying optimizations to enable the chain to reach one million transactions per second, rather than simply building a separate side chain for each application.
What are the advantages of Solayer's hybrid consensus mechanism over POW & POS?
Jason: You can think of Solayer as a product similar to Arbitrum on Ethereum. But Arbitrum's consensus confirmation mechanism on Ethereum first provides an excellent user experience - transactions can be quickly confirmed after users click, and many large transfers and DeFi operations are run on Arbitrum. Our idea is that we do not want to adopt a pre-confirmation-based system, but a voting-based system. The specific approach is very simple, which is to transplant this voting-based mechanism to a chain similar to the Arbitrum architecture. Once this is done, it actually becomes an independent L1, not an L2. The difference between L2 and L1 is whether the root state of the chain is affected by the main chain. For example, the root state of Arbitrum is anchored to Ethereum, but we are not restricted by this.
The main purpose of our work is to make the trust mechanism no longer dependent on the verification mechanism, but more active, rather than passive like Arbitrum. The so-called passive means that users need to run nodes every day to verify potential errors, and this verification is usually extremely low-return and only applies to very small probability situations. Therefore, we chose to change this and allow users to directly actively verify and vote, while we do not want to completely decentralize block construction. For example, in Solana, different validators are switched every 400 milliseconds, which raises some issues, such as the multi-validator mechanism (MVV) may cause problems in block construction because for end users, the order of transactions does not have to be determined by multiple different participants. If it is only for the purpose of achieving proof of censorship resistance, block construction can be designed to be more trust-minimized in nature.
Therefore, we adopted a model based on authorization proof, that is, a selected node is responsible for block construction; when the node fails, it can fall back to other main chains for election, and other nodes are mainly responsible for providing weight to determine the chain status. This is the advantage of our hybrid consensus mechanism over POW and POS.
What was the key factor for Solayer to receive investment from Polychain and Binance Labs?
Jason: Although financing is not within my scope of responsibility, I can explain it briefly. I think it is very important to have a long-term path planning, which is a key factor. Secondly, the cycle is also important, and the timing is also very critical. No matter what you do, timing is crucial.
What does Solayer’s current product matrix include?
Jason: We have a yield layer product, and another one is sUSD, which is similar to treasury bonds, such as US treasury bonds; in addition, SOL staking is a Solana liquid staking solution, and although the yield is not that high, the risk is not great at the same time. This is what we do. And an internationally available crypto card. If you deposit USDC into this card, you can spend it directly through Visa, and everyone will be able to start applying soon. The ultimate vision of this product is to use InfiniSVM to support all transactions.
Where will Solayer be headed in 2025?
Jason: I think it is mainly from a technical perspective that it is very interesting to be the first chain to reach 1 million TPS and to allow everyone to test it online. Apart from that, we don’t have many other ideas. The team has been trying many things, such as trying products such as U cards, and doing more C-end applications, but the most important thing is how to become the fastest chain, which is the most important thing for us.
