Primus Labs: Infrastructure Connecting Verifiable Data with Confidential Finance The next step in Web3: Trustworthy data and secure computation Blockchains have created a transparent environment where anyone can verify transactions. However, bringing real-world data onto the blockchain and confirming its accuracy remains challenging. Consider data such as exchange asset holdings, user financial information, or data generated by web services. While valuable for blockchain applications, publicly exposing the original data risks privacy and security breaches. Primus Labs is developing cryptographic infrastructure to solve precisely this problem. At its core are technologies like zkTLS, FHE, and zkVM—enabling verification of data integrity while protecting sensitive information and performing necessary computations. 1. zkTLS: Verifying the Trustworthiness of Web Data Data stored in Web2 services holds significant value for Web3 applications. Yet, merely claiming that data was retrieved from a specific website is insufficient to prove its authenticity. Primus leverages zkTLS—a technology designed to verify the origin and validity of web data based on HTTPS communications. This allows users to prove facts—such as holding a specific balance or financial condition in an exchange account—without revealing the full original dataset. Crucially, zkTLS is not a tool for simply moving data onto the blockchain. It enables verification that the data originated from a legitimate service through direct communication with it. 2. Combining Data Verification and Computation: DVC Verifying data origins alone is not enough. It must also be possible to compute specific conditions based on verified data and prove that the results were correctly derived. Primus’s DVC (Data Verification and Computation) is an architecture that bridges these two processes. While zkTLS verifies the source and authenticity of data, zkVM can prove that computations performed on that data were executed correctly. TEEs provide isolated execution environments that protect data during computation. For example, suppose a user wants to prove they hold above a certain asset threshold in their exchange account. Instead of disclosing their full balance, they can present only a verifiable proof that the condition has been met. This approach has broad potential applications in financial services, identity verification, on-chain reputation systems, and more. 3. FHE: Performing Computations on Encrypted Data Another key technology Primus focuses on is FHE (Fully Homomorphic Encryption). Traditional encryption methods typically require data to be decrypted before any computation can occur. FHE, however, enables computations to be performed directly on encrypted data without decryption. This is especially critical for services handling sensitive financial information: providers can perform necessary calculations without ever accessing the raw data. Importantly, FHE does not serve the same role as zkTLS or zkVM. Each technology addresses distinct challenges; when used together, they provide broader support for both data verification and privacy protection. 4. AutoHoG: Enhancing FHE Performance The ability to compute on encrypted data is a major advantage—but FHE can demand significantly higher computational resources than conventional operations. To address these performance limitations, Primus is also researching GPU-accelerated solutions such as AutoHoG. According to project documentation, AutoHoG achieved approximately 90ms latency per 64-bit transmission and over 210 operations per second using an environment powered by eight NVIDIA RTX 5090 GPUs. It is important to note that these figures represent internal test results from the project and should be distinguished from independently verified, universal performance benchmarks. Nevertheless, efforts to accelerate encrypted computation are essential for enabling FHE’s practical adoption in real-world services. 5. From Research to Real-World Applications: Proof of Reserves and Confidential Finance Primus’s technologies are not confined to research. Proof of Reserves can be implemented using Primus’s stack: zkTLS verifies the trustworthiness of external data sources, followed by secure computation and proof generation to demonstrate reserve holdings. Another key application is Confidential Finance. Primus is expanding services on the BNB Chain ecosystem using FHE to protect sensitive financial details such as account balances and transaction amounts. One critical distinction: Confidential Finance does not fully conceal all transaction information. Depending on implementation, certain details—such as wallet addresses or the fact that a transaction occurred—may still remain visible on-chain.Therefore, it is more accurate to view this not as equivalent to complete anonymity or a mixer, but as an approach aimed at reducing the exposure of financial information within a public blockchain environment. Why is Primus’s technology important? For Web3 to scale to more users and real-world financial services, it requires technologies that go beyond merely recording transactions. Real-world data must be trustworthy, and it must be possible to verify that computations performed using this data have been executed correctly—while ensuring users’ sensitive information is not unnecessarily exposed. The technologies developed by Primus Labs—zkTLS, DVC, FHE, and zkVM—are attempts to address these requirements through distinct approaches. Performance and implementation challenges remain unresolved. However, the direction of integrating data verification, verifiable computation, and privacy protection within a single infrastructure has the potential to play a crucial role in the future of Web3 finance. Ultimately, Primus’s core mission is not simply to bring data onto the blockchain—it is to prove that data originates from trustworthy sources and enable necessary operations while safeguarding sensitive information. This is why Primus Labs should be viewed not merely as a blockchain project, but as an infrastructure initiative addressing the trust gap between Web2 and Web3. @primus_labs