The crypto industry has united: AI labs must open access to their models to protect assets
In recent weeks, it has become clear that defenders of digital infrastructure are left without the most powerful tool in their arsenal — advanced artificial intelligence models. I have carefully studied the situation and come to the conclusion that this is a systemic problem that requires immediate intervention at the highest level.
My analysis shows that the open letter initiated by the Bitcoin Policy Institute and supported by dozens of leading players — from Coinbase and Block to BitGo and Trezor — is not just a request, but an ultimatum. It is about providing verified open-source developers with reliable access to the strongest AI models. Without this, we will not be able to adequately counter increasingly sophisticated attacks on an ecosystem where trillions of dollars are at stake.
Critical vulnerability amid trillion-dollar risks
Digital assets have long outgrown the status of a speculative instrument. The Bitcoin network protects more than a trillion dollars, and the entire ecosystem — from wallets and cryptographic libraries to custodial systems and exchanges — manages sums comparable to the GDP of an average state. Behind this code stand the personal savings of millions of people, pension funds, and reserve funds. A single vulnerability can destroy the wealth of entire generations, so the appeal for cyberattacks is obvious.
The situation is compounded by the fact that attackers are already actively using AI. According to my data, experienced hackers, likely from other jurisdictions, are conducting continuous automated attacks. Small teams of open-source maintainers, including Bitcoin Core developers, physically cannot keep up with this pace, having only limited access to computing resources.
Imbalance of power: defenders at a disadvantage
The key problem is that leading laboratories provide access to their models only to select partners, leaving independent researchers out in the cold. Public interfaces often block legitimate security work, forcing defenders to settle for less powerful open-weight alternatives. This puts them in a deliberately losing position: while defenders analyze code manually, attackers automate vulnerability discovery with AI.
The coalition demands not just access, but the creation of structured trusted access programs. This includes early controlled access to the strongest models, sufficient compute budgets for autonomous agents, and secure environments for analyzing closed-source code. Particular emphasis is placed on including small non-profit organizations on par with large corporations and states.
The industry is ready to take on coordination: setting standards, verifying participants, and ensuring responsible vulnerability disclosure. It is my firm belief that AI can become the most powerful defensive technology in history, but only with fair access. If laboratories continue to ignore this call, we risk a digital collapse that could have been avoided.
My verdict: this is not a matter of convenience, but a matter of ecosystem survival. Refusing dialogue will cost more than any risks associated with opening access.