Claude Cracked a $400K Bitcoin Wallet , What That Means for Security Teams
Anthropic's Claude iterated through 3.5 trillion password candidates to recover an 11-year-old wallet, reframing LLMs as serious cryptographic recovery tools.
10. Claude Cracked a $400K Bitcoin Wallet , What That Means for Security Teams
A Bitcoin holder who lost access to a wallet in 2014 recovered roughly $400,000 worth of BTC after using Anthropic's Claude to orchestrate a brute-force password recovery campaign. The AI worked through an estimated 3.5 trillion password candidates against an old wallet backup before successfully decrypting it. The recovery was reported by Tom's Hardware in May 2026 and drew 242 upvotes on Hacker News, a reliable signal that security practitioners are paying close attention.
The strategic implication sits at the intersection of two markets: AI-assisted security tooling and the still-fragmented crypto wallet recovery space. Dedicated recovery services like Dave Bitcoin and Wallet Recovery Services have operated for years, charging 20-30% of recovered funds. Claude doing this work at the direction of an end user, without a specialist intermediary, puts direct pressure on that fee model. More broadly, it confirms what red teams have suspected: LLMs are not just reasoning assistants but capable orchestration layers for compute-intensive, rule-based enumeration tasks. That changes how defenders should think about password hygiene and key derivation schemes, particularly for anything generated before 2015 with weaker entropy assumptions.
The pattern worth watching is whether Anthropic positions this capability deliberately or treats it as incidental. OpenAI and Google DeepMind have both been careful to gate or disclaim security-adjacent use cases. If Anthropic stays quiet, the community will fill the gap with informal tooling built on top of the Claude API, which creates compliance exposure. The next signal to track: whether this surfaces in enterprise security vendor partnerships, or whether regulators in the EU or UK flag LLM-assisted cryptographic attacks as a category requiring disclosure frameworks.
Source: Bitcoin trader recovers $400,000 using Claude AI after losing wallet password 11 years ago