Over 40 Crypto Firms Demand AI Labs Grant Security Researchers Model Access

More than 40 cryptocurrency and bitcoin companies have called on the world’s leading artificial intelligence labs to grant independent security researchers access to their most advanced models before they become publicly available.

The push comes through an open letter published on August 10, organized by the Bitcoin Policy Institute. Prominent signatories include Coinbase, Block, BitGo, Blockstream, and ARK Invest, alongside dozens of other firms defending the infrastructure that underpins the wider crypto ecosystem.

The Asymmetry Problem

The letter rests on a straightforward concern: there is currently an uneven playing field when it comes to who can access the most powerful AI capabilities. The Bitcoin network secures more than a trillion dollars in value, with additional trillions locked across wallets, custody systems, signing devices, and supporting infrastructure built on open-source code. A single vulnerability in this code can result in permanent loss of funds for users.

Despite this critical role, the developers and maintainers responsible for keeping this infrastructure secure cannot access the vetted researcher programs that AI labs operate internally. When these developers attempt to use publicly released models to identify vulnerabilities, the built-in safety guardrails often prevent the kind of legitimate security research needed to find flaws before attackers exploit them. As a workaround, many resort to open-weight models they can run independently—effective tools, but materially less capable than the frontier systems that malicious actors already access or will access when models enter the public domain.

A Call for Trusted Access Programs

The open letter asks frontier AI labs to establish or expand standing trusted-access programs for qualified defenders of open-source financial infrastructure. The signatories argue that early access to advanced models and reasonable compute budgets would directly help core developers write more secure code and identify vulnerabilities before they cause real losses.

The letter highlights a key reality: AI labs and their partners already conduct months of testing and evaluation of new attack capabilities before any public release. These capabilities inevitably escape into the broader ecosystem anyway through public releases and through compromised systems. The companies reason that letting qualified defenders access those same capabilities in advance could shift the security equation meaningfully in their favor.

Security at Scale

The move reflects growing urgency around crypto security. Last month, the Bitcoin Security Consortium—a coalition including BlackRock, Coinbase, Anchorage, ARK, Block, Blockstream, Fidelity Digital Assets, and Galaxy—committed $15 million specifically to long-term network security. The new open letter targets a different but complementary problem: not funding tomorrow’s solutions, but enabling today’s defenders to use the most capable tools already available.

So far, major AI labs have not publicly responded to the specific request, though they have indicated work on safeguards designed to prevent their models from being misused—whether for blockchain-related attacks or other malicious purposes. Security researchers have documented cases where AI-assisted attacks have powered live Bitcoin ransom campaigns, making clear that the tools available to attackers continue to evolve. Granting defenders equivalent access means carefully vetting who qualifies and trusting them not to leak capabilities—a risk labs have been reluctant to take at scale.

Stronger AI-powered security for bitcoin infrastructure tightens defenses across the entire crypto ecosystem, potentially reducing systemic vulnerabilities that threaten user assets across digital finance.

Source: Bitcoin Policy Institute, via Decrypt. Not financial advice.