Solana sounds the alarm: AI can break post-quantum cryptography before quantum computers do

Solana’s co-founder questions the future security of blockchains: according to Anatoly Yakovenko, artificial intelligence (AI) could discover flaws in post-quantum systems even before quantum computers become a reality.

This position completely shifts the focus of the debate. In recent years, the sector has in fact focused on so-called “post-quantum cryptography”, that is, algorithms designed to withstand future quantum attacks.

However, Yakovenko suggests that the problem could emerge much earlier, through tools that are already available today.

The paradox of post-quantum security and its link with AI

As we know, blockchains are preparing for the fact that quantum computers could render current digital signature systems obsolete.

Precisely for this reason, Solana has started considering Falcon signatures, a scheme regarded as one of the most promising candidates for post-quantum security.

The idea is simple: progressively replace current cryptographic mechanisms with more robust versions. However, the reality is much more complex. Yakovenko points out that we do not yet know all the possible vulnerabilities of these new systems.

And this is true not only at the mathematical level, but also in practical implementation. In this context, AI plays a fundamental role, since there are already models capable of analyzing code, identifying patterns, and discovering errors that escape humans.

In other words, algorithms that are considered secure today could be challenged much earlier than expected.

Unsurprisingly, the core of the warning concerns precisely the role of AI. If in the past breaking a cryptographic system required years of research and limited resources, today automated analysis tools drastically accelerate this process.

Yakovenko is not claiming that post-quantum signatures are already vulnerable, but he highlights a systemic risk. This means that the industry could adopt new standards without having fully understood all the possible modes of failure.

This is therefore a radical change in the way we think about security, since it is no longer just about ‘defending against a specific enemy’, such as the quantum computer, but about facing an environment in which attack capabilities are constantly evolving.

Solana’s response: more schemes, less dependence

In the face of this uncertainty, Yakovenko has proposed avoiding dependence on a single cryptographic scheme. Instead of relying entirely on one post-quantum solution, he suggests an approach based on multiple layers of security.

The idea is to use two or three different signature schemes, creating a sort of cryptographic redundancy. In this way, even if one of the systems turns out to be vulnerable, the others could still guarantee security.

This approach recalls the concept of “defense in depth”, already used in other areas of cybersecurity. However, applying it to blockchain entails significant technical challenges, especially in terms of performance and complexity.

Within this scenario, as mentioned, Falcon signatures are one of the most discussed elements. Developed to be efficient and compact, they are considered suitable for high-throughput blockchains like Solana.

As far as we know today, developers are working to optimize performance, reducing the computational cost of verifications.

This is a crucial step, because any post-quantum solution must be compatible with the operational needs of existing networks.

However, Yakovenko urges caution. Even if Falcon or other similar schemes perform well in tests, there is no guarantee that they are immune to vulnerabilities that may be discovered in the future.

The debate in the crypto community

In any case, Yakovenko’s statements have sparked discussion among developers and researchers. Some believe that formal verification of systems can significantly reduce risks, ensuring a high level of security.

Others, instead, share the concern that it is not possible to foresee all potential weaknesses. Cryptography is a complex field, and every new solution introduces variables that are difficult to fully control.

Moreover, this debate reflects a broader tension in the sector. On the one hand, there is the need to innovate and prepare for the future. On the other, the risk of adopting solutions that are still immature.

Not only that, Solana’s warning also fits into a broader context. In recent months, in fact, several blockchains have begun to explore post-quantum solutions, while others are evaluating alternative approaches.

At the same time, artificial intelligence is becoming increasingly central in the field of security as well. Not only as a threat, but also as a defensive tool, capable of identifying vulnerabilities before they are exploited.

This dual nature of AI makes the picture even more complex, since the same technology that can break a system can also help strengthen it.

Timing certainly plays a central role. Not by chance, there are several experts who believe that quantum computers capable of compromising cryptography are still far from being developed.

Yakovenko, instead, suggests that the risk related to AI is immediate. This changes the priorities for developers, who must decide where to focus their resources.

Investing only in protection against quantum threats may therefore not be enough. A broader vision is needed, one that takes into account the threats that exist today.