Major firms including Coinbase, Block, and BitGo have signed a letter stating that current safety measures on AI models hinder legitimate security efforts while attackers operate without such restrictions.
By Shaurya MalwaUpdated 21 min agoPublished 39 min ago2 min readMake preferred on ShareShare this articleCopy linkX (Twitter)LinkedInFacebookEmailMake preferred on Bitcoin companies request AI labs to provide tools for security research. (Kevin Ku/Unsplash/Modified by CoinDesk)SummaryShow- Over 30 cryptocurrency firms are calling on leading AI labs to grant open-source security researchers early access to their top models, claiming that defenders are at a disadvantage with inferior tools compared to attackers.
- The letter, initiated by the Bitcoin Policy Institute and endorsed by firms like Coinbase, Block, and ARK Invest, highlights that Bitcoin Core developers are excluded from trusted partner programs and face limitations from safety filters on public models.
- Recent security breaches involving BTCPay Server and Lightning nodes, identified partly by an AI-driven volunteer group known as the “Bitcoin Red Team,” strengthen the signatories’ request for early access to models, computational resources, secure testing environments, and direct communication with lab security teams.
More than 30 companies in the Bitcoin and cryptocurrency sector have reached out to major AI labs, requesting that they provide open-source security researchers with early access to their most advanced models. They argue that those tasked with defending a trillion-dollar ecosystem are equipped with less effective tools compared to those launching attacks.
The letter, organized by the Bitcoin Policy Institute and released earlier this week, boasts signatures from Coinbase, Block, BitGo, Blockstream, Anchorage Digital, ARK Invest, Bitwise, Foundry, Casa, Exodus, and various nonprofit developer funds such as Brink, Chaincode, and Btrust.
The primary concern raised in the letter is that Bitcoin Core developers, the small team responsible for maintaining the network's software, are unable to access programs offered to trusted security partners.
When they resort to publicly available models, they encounter safety filters that are intended to prevent the creation of malware, which inadvertently obstruct efforts to identify vulnerabilities ahead of malicious actors.
This leaves them dependent on open-weight models, which are less powerful and freely downloadable.
In contrast, attackers do not face similar limitations. The letter asserts that select labs and their partners can access new offensive capabilities months in advance, while those capabilities eventually spread through public models, unauthorized access to corporate systems, and custom hacking tools.
The signatories are requesting five specific actions: early access to the most capable cyber models, computational resources for thorough evaluations, secure environments for reviewing private code, inclusion for small and independent maintainers in addition to large firms, and direct communication lines with lab security teams for reporting findings.
The timing of this request is significant. Two of the organizations that signed the letter have recently highlighted both sides of this issue.
BTCPay Server, one of the signatories, reported a major vulnerability last week that attackers had already exploited to siphon funds from merchants' Lightning nodes. Foundation, a hardware wallet manufacturer that also signed the letter, lost one of its own nodes in this attack. Following this incident, BTCPay noted that the evolution of AI is shifting the dynamics between attackers and defenders, as AI models enable quicker and more cost-effective searches through extensive codebases for vulnerabilities.
It was defenders who identified the vulnerability. A volunteer collective known as the Bitcoin Red Team began leveraging AI models to analyze Bitcoin codebases this month and has submitted thousands of findings across numerous projects, including the report that led to the patch for BTCPay.
Related AssetsBitcoin$63 836,560.049%Latest Crypto News- 1Metaplanet denies selling bitcoin worth $320 million 25 min ago
- 2Speculation on dogecoin is back to October 2025 levels. The price is down 70%1 hr ago
- 3Bitcoin slips near $63,500 as traders look past CPI to Fed’s next tests2 hrs ago
- 4Prediction markets should dial back faulty filings for incentives to boost trading: CFTC6 hrs ago
- 5Securitize falls 20% after earnings miss as tokenization revenue falls short7 hrs ago
- 6Morgan Stanley's infrastructure partner Zerohash rebuffed in pitch to be U.S. trust bank7 hrs ago
- 7Goldman Sachs leaps into bitcoin income ETFs with $2.25 billion NEOS buyout12 hrs ago
- 8Wintermute plans $1 billion AI push beyond crypto: Bloomberg13 hrs ago
- 9Kalshi taps DoubleZero for Wall Street-style high-speed data feed13 hrs ago
- 10Miden bets on privacy stablecoins with introduction of USDCx15 hrs ago
Building the Zcash Machine: Tachyon and Quantum Readiness
Building the Zcash Machine: Tachyon and Quantum Readiness
Zcash’s Tachyon upgrade aims to scale shielded payments, improve quantum readiness, and test whether its funding, security, and governance can hold.
By CoinDesk ResearchJun 30, 2026Commissioned byGenZcashZcash’s Tachyon upgrade aims to scale shielded payments, improve quantum readiness, and test whether its funding, security, and governance can hold.
Why it matters:
Zcash’s Tachyon upgrade aims to scale shielded payments, improve quantum readiness, and test whether its funding, security, and governance can hold.
View Full ReportMore From Tech