LOS ANGELES, Nov. 26, 2025 /PRNewswire/ — In a breakthrough that combines privacy, cryptographic integrity and intelligent automation, a purpose-built AI model for smart contract development has been successfully deployed inside a confidential virtual machine on Secret Network. Solidity-LLM, a large language model developed by ChainGPT, is now running within SecretVM, the Confidential Virtual Machine framework that underpins Secret Network’s latest evolution in decentralised computing.
This is the first time an AI model trained exclusively to write and audit Solidity code has been embedded into a Trusted Execution Environment. The implications for blockchain development are significant. Developers can now leverage artificial intelligence to generate, optimise and analyse smart contracts without ever exposing the source code or proprietary logic to a third party. Privacy is preserved by design, and verification is built into the cryptographic structure of the system.
At the heart of this development lies SecretVM, a privacy-preserving compute layer built atop hardware-based TEEs, including Intel’s TDX and AMD’s SEV. SecretVM offers confidentiality, execution integrity and remote attestation in a single runtime. It allows sensitive workloads, such as AI inference, financial computations or cross-chain coordination, to run inside a sealed enclave. In practical terms, even the node operator cannot access the data, inputs or outputs processed by the AI.
“Confidential computing is no longer an abstract concept,” said Luke Bowman, COO of the Secret Network Foundation. “We’ve shown that you can run a complex AI model, purpose-built for Solidity, inside a fully encrypted environment and that every inference can be verified on-chain. This is a real milestone for both privacy and decentralised infrastructure.”
Solidity-LLM was developed by ChainGPT to address the specific needs of Ethereum and EVM-compatible smart contract development. It is a 2-billion-parameter model trained on over 650,000 curated Solidity contracts. This domain-specific approach gives the model a nuanced understanding of contract logic, optimisation patterns and security practices that general language models cannot replicate. Now, running within SecretVM, the model operates entirely in private, enabling confidential, on-chain software development for the first time.
Christopher Duggan, Head of Marketing, ChainGPT said, “We’ve spent the last year building and fine-tuning a model that understands the real structure of Solidity. What makes this deployment so important is that developers can now use the model without sacrificing their IP or exposing critical business logic. Everything stays encrypted and verifiable.”
This deployment solves a long-standing paradox in AI-assisted development. Traditional coding assistants require developers to upload source code to a centralised cloud platform. In doing so, they risk exposing intellectual property, user data and business logic. Within SecretVM, however, Solidity-LLM performs inference in an environment that is both encrypted and verifiable. Cryptographic proofs ensure the integrity of the enclave, meaning developers and institutions can trust not only the AI’s output but also the infrastructure within which it operates.
The architecture supporting this deployment is both robust and flexible. Solidity-LLM runs as a containerised workload inside SecretVM, leveraging Docker to support diverse language frameworks and integration methods. Access is provided via API, SDK or direct smart contract interfaces. Developers can treat the model as a programmable agent in its own right, capable of interacting with contracts, tools and governance frameworks, all without revealing sensitive inputs.
The use cases are far-reaching. For developers and builders, it means confidential co-development. Code can be generated, tested and optimised in private, with no fear of leaking proprietary logic. For auditors and security professionals, the system offers a trusted environment to review and analyse contracts with AI assistance. No data leaves the enclave, maintaining confidentiality throughout the audit process.
For institutions and enterprises, this architecture provides a clear path to blockchain automation without the compliance and privacy risks associated with conventional AI tools. Smart contracts can now be generated and deployed in a way that aligns with regulatory standards, particularly in sensitive sectors such as finance, healthcare or governance.
The implications for decentralised finance and DAOs are especially compelling. Smart contract agents can now operate autonomously inside a cryptographically verified environment. They can manage upgrades, execute governance decisions or coordinate cross-chain logic, all without disclosing internal states or decision-making data.
Researchers and AI engineers also benefit. SecretVM supports federated training and fine-tuning across encrypted datasets. This allows for collaborative AI development without compromising data sovereignty. It enables secure multi-party computation, where models evolve in tandem with decentralised networks rather than in corporate silos.
This integration shifts the trust model for AI on-chain. It moves from opaque clouds to verifiable compute, from public exposure to private collaboration, and from centralised inference to decentralised autonomy. Developers retain ownership of their data, their code and their infrastructure.
The roadmap includes confidential model fine-tuning, orchestration of multiple AI agents and cross-chain deployments. As artificial intelligence continues to merge with blockchain infrastructure, the underlying challenge remains the same – how to build trust without giving up control. This deployment offers a practical and proven answer.
View original content to download multimedia:https://www.prnewswire.com/news-releases/private-by-design-secretvm-and-solidity-llm-usher-in-confidential-ai-for-web3-302627075.html
