Over 40 crypto firms warn on AI

TLDR

More than 40 major Bitcoin and crypto firms have signed an open letter warning that current AI policies are giving attackers an edge over defenders securing digital asset infrastructure.

1. The coalition asks leading AI labs to grant vetted security researchers early, trusted access to their strongest models so wallets, nodes, and payment software can be audited before exploits happen.
2. They argue that today, attackers and select partners already use frontier AI to find bugs, while open source maintainers protecting trillions of dollars in crypto are restricted to weaker tools.
3. Next to watch: whether AI labs create defenders-first access programs and how regulators, such as the CFTC, fold AI-assisted security into their emerging crypto rulebooks.

Deep Dive

1. What The Firms Are Warning About

Over 40 Bitcoin and digital asset organizations, including Coinbase, BitGo, Block, Blockstream, Anchorage Digital, ARK Invest, Bitwise, Galaxy, Trezor, Brink and Chaincode Labs, signed an August 10 open letter coordinated by the Bitcoin Policy Institute and titled Defenders Need the Frontier here.

The letter asks leading AI labs to give vetted open source security teams controlled early access to frontier models, plus adequate compute, secure review environments and direct lines to lab security staff.

They stress they are not asking for unrestricted public release, but for structured trusted access so small nonprofits and independent developers can realistically defend Bitcoins trillion dollar stack and other critical crypto software.

2. Why This Matters For Crypto Security

Recent incidents underline the concern. The letter cites a Coldcard hardware wallet breach and a critical BTCPay Server flaw that allowed attackers to steal Lightning Network funds, both showing how a single bug can vaporize user savings.

A volunteer Bitcoin Red Team has already used advanced models like Moonshots Kimi K3, alongside OpenAI and Anthropic, to scan 390 projects in under 28 hours, generating thousands of findings, including dozens of critical vulnerabilities, demonstrating how powerful AI can be as a defensive tool if defenders can access it Bitcoin Red Team audit.

The signers warn that today, attackers and privileged commercial partners often get months of early access to new capabilities, while open source maintainers face safety guardrails that block realistic security work, widening a security divide.

3. What To Watch Next

So far, AI labs have not publicly committed to the kind of defenders-first access program the letter outlines, but some have acknowledged that the same models can both harden and attack crypto systems.

Regulators are starting to put AI and crypto on the same agenda. The CFTCs Innovation Advisory Committee will discuss crypto assets and AI together at its inaugural meeting, signaling that AI-assisted security and market abuse are entering formal policy conversations CFTC crypto and AI rules discussion.

Projects that move quickly to integrate AI-assisted audits, formal red teaming, and transparent vulnerability disclosure pipelines are likely to be better positioned if labs, regulators, or insurers begin to treat such practices as baseline expectations.

Conclusion

This warning from over 40 crypto firms highlights a new security reality: frontier AI has already changed the offensedefense balance for digital assets. If AI labs, regulators, and open source maintainers can align on trusted access and robust audit practices, AI could become a core pillar of crypto security rather than a one-sided weapon in attackers hands.