More than 40 bitcoin and digital-asset organizations have signed an open letter asking the world's leading AI laboratories to give vetted security researchers controlled access to frontier models, arguing that the developers who protect Bitcoin's infrastructure are falling behind the attackers targeting it. The letter, titled "Defenders Need the Frontier" and organized by the Bitcoin Policy Institute, was published on August 10.
Signatories span custodians, exchanges, mining firms and open-source developer funds, including Coinbase, Block, BitGo, Blockstream, Anchorage Digital, ARK Invest, Bitwise, Foundry, Strategy, MARA, Galaxy and Trezor, alongside nonprofit groups such as Brink, Chaincode Labs, Btrust, OpenSats and BTCPay Server. The letter also invokes the Bitcoin Security Consortium — a group including BlackRock, Coinbase, Strategy, Anchorage, ARK, Block, Blockstream, Fidelity Digital Assets and Galaxy that pledged $15 million toward Bitcoin's long-term security in late July.
An asymmetry the letter says favors attackers
The letter's central argument is one of asymmetry: Bitcoin alone secures more than $1 trillion in value, with trillions more held across wallets and custody systems built largely by small teams and volunteer maintainers. Those defenders, the letter says, often can't get access to the most capable AI systems used for cybersecurity research, while safety guardrails on publicly available models can shut down legitimate vulnerability hunts before they start — even as sophisticated attackers already operate with fewer such restrictions.
The coalition is explicit that it isn't asking for unrestricted public access to models with advanced cybersecurity capabilities. Instead it wants labs to establish or expand “standing trusted-access programs” that would give qualified defenders early access to frontier cybersecurity models, sufficient compute, secure research environments and direct lines to lab security teams. Notably, the letter asks labs not to limit eligibility to large companies and governments, pointing out that small nonprofits and independent maintainers protect software used by millions despite lacking the institutional ties needed to join existing programs.
The threat the letter is responding to
The urgency traces partly back to Anthropic's own disclosures about AI-assisted crime. In a report the company published, a cybercriminal used Claude Code to run a mass extortion campaign against at least 17 organizations across government, healthcare and religious sectors, with the AI model automating reconnaissance, credential harvesting and network penetration, and even drafting ransom notes demanding between $75,000 and $500,000 — paid in Bitcoin.
Related: Whale Loses $26M in Private Key Breach, Two Years After a $24M Phishing Hit
That case, widely cited in security circles as the first fully AI-driven cyber campaign to succeed in the wild, is the kind of scenario the new letter says defenders need equivalent tools to counter.
None of the major AI labs had publicly committed to the coalition's request as of publication. The letter remains open for additional signatures, and its authors say bitcoin users and developers should watch for lab responses, new audit results and further disclosures from open-source projects that lean on AI-assisted security reviews.
