In collaboration with Grok
The code analysis and vulnerability discovery capabilities of artificial intelligence have far surpassed those of humans, reaching thousands of times greater. Many deep-seated issues that once relied on manual auditing are now precisely and efficiently identified by AI. More critically, AI capabilities are advancing at a monthly pace: vulnerabilities undetectable today may be exposed and automatically exploited tomorrow. The speed of code auditing has clearly lagged behind AI development. Consequently, an open-source project considered secure today may face new risks in the future.
This means that the security of open-source code is no longer a static certainty, but a dynamic uncertainty.
This uncertainty affects different types of projects in markedly different ways.
For open-source AI models, the issue is relatively manageable, as models are tools whose uncertainty is an inherent attribute and does not directly pose systemic financial risks. However, in financial systems that demand high certainty, projects lacking such certainty are unacceptable. Cryptocurrencies fall precisely into this category of financial projects and will thus become the hardest hit by security shocks in the AI era. According to Chainalysis reports, direct losses from hacker attacks on cryptocurrencies in 2025 exceeded $3.4 billion. The automated vulnerability discovery and attack capabilities driven by AI will undoubtedly further amplify this risk, empowering attackers even more.
Bitcoin: Core Stability Tested by 17 Years of Real-World Operation
Bitcoin has been stably running for nearly 17 years since the genesis block in January 2009. Its core code is extremely concise and has undergone extensive real-world testing, offering relatively high reliability. However, recent ecosystem expansion practices (such as Ordinals inscriptions and Runes token protocols based on SegWit and Taproot) have to some extent increased systemic uncertainty. While these innovations have brought substantial fee revenue to miners, some community members view them as “junk data” occupying block space, sparking stability debates. The BIP-110 proposal put forward at the end of 2025 (aimed at restricting non-monetary data through a soft fork) has further intensified this discussion, which continues to provoke widespread debate within the community.
Bitcoin cannot abandon the open-source model, but it must exercise extreme caution in code extensions, prioritizing the protection of its core function as a store of value to minimize uncertainty to the greatest extent.
Ethereum: The Dilemma of Functional Expansion and AI Integration
Ethereum also has a history of about 12 years, with its code having undergone long-term validation. It successfully completed the major transition from PoW to PoS in September 2022 (The Merge). At that time, the rise of the Bitcoin ecosystem was partly driven by market concerns over Ethereum’s stability. Fortunately, the transition proceeded smoothly without major issues.
However, Ethereum’s functional expansion continues. The two major upgrades in the 2026 roadmap—Glamsterdam (first half) and Hegotá (second half)—focus on improving execution efficiency, proposer-builder separation (ePBS), Verkle tree state optimization, and actively integrating artificial intelligence tools. Core figures such as Vitalik Buterin have repeatedly emphasized Ethereum’s AI vision, including trustless AI interactions, AI agent economic infrastructure, and governance systems.
Failing to integrate AI makes it difficult to remain competitive, yet integrating AI introduces new security uncertainties, creating a classic dilemma. To address this challenge, project teams should extend testnet operation periods, strictly control code increments, and routinely incorporate AI-assisted security audits.
Application-Layer Projects: High Risks Under Frequent Iteration
Application-layer projects involve more frequent code improvements and upgrades, with larger changes and more potential vulnerabilities. The traditional assumption that “open source equals security” faces greater challenges here. It is recommended to adopt a non-open-source or partially open-source strategy, while ensuring that the ledger (transaction history) and key states remain publicly transparent to maintain auditability and minimize trust.
Transparent Centralization + Public Ledger: A Pragmatic Trade-off Solution in the AI Era
From the above analysis, in the AI era, “open source” is no longer an absolute golden rule for cryptocurrency projects. The technological conditions of the market have undergone fundamental changes—AI has accelerated vulnerability discovery from “manual monthly audits” to “machine real-time detection,” rapidly eroding the security benefits of the traditional open-source model.
Transparent centralization combined with a public ledger may become one pragmatic trade-off solution for addressing security challenges in the AI era. The Hyperliquid project serves as a typical example: its core node/validator code remains closed-source (with the team stating it will gradually open source once the system is sufficiently secure and stable), while the SDK and some tools are public, which has also sparked ongoing community discussions regarding the degree of decentralization. However, its transaction ledger is fully on-chain and publicly transparent. This model achieves extreme execution speed while providing stronger certainty guarantees under AI-driven security threats.
Of course, this model is not perfect and still requires further strengthening through multi-layered mechanisms such as formal verification, zero-knowledge proofs, and extended testnet operation. In the future, cryptocurrency projects are likely to widely adopt hybrid architectures: the core consensus layer should be opened cautiously, while the application layer can remain closed-source but supplemented with verifiable execution environments, striking a balance between innovation speed and financial-grade certainty.
Conclusion
The AI era has fundamentally altered the security paradigm of open-source code. Cryptocurrency projects must proactively adapt—replacing dogmatic open-source approaches with more pragmatic transparency designs—to fortify the foundation of financial systems amid uncertainty. Bitcoin’s cautious expansion, Ethereum’s balanced innovation, and Hyperliquid-style transparent centralization practices are vivid examples of the industry exploring new paths under the impact of AI. In the future, those who can better manage this “uncertainty” will gain the upper hand in the next cycle.
