Bitcoin infrastructure is facing heightened security concerns as artificial intelligence makes it cheaper to identify vulnerabilities in complex financial software. Reports said more than 5,200 Bitcoin addresses were hacked this month, with losses exceeding 1,800 BTC, while attackers drained about $114 million from Coldcard wallets and white-hat hackers extracted roughly 4,000 BTC, valued at $317 million, from Blockstream's Liquid Network before returning 3,400 BTC after a patch. Bitcoin Core developers also discussed the project's AI readiness at a weekly meeting, after Spiral head Steve Lee said donors and community members feared it was underprepared for AI-related network threats and was not using AI effectively for code review, testing and troubleshooting. Hackers reportedly used Kimi K3 to identify a Coldcard vulnerability, while Core Lightning developers issued an emergency fix after AI-generated reports found genuine flaws. In August, 16 Bitcoin developers used AI models to examine 390 projects and generated nearly 5,000 findings, including 85 initially classified as critical. The incidents reflect a design trade-off: Bitcoin's main layer remains intentionally simple to reduce risk, while smart contracts and off-chain scaling layers add functionality through more complex codebases. Gregory, a bitcoin application developer and CommerceBlock co-founder and CEO, said AI is finding bugs humans may miss and questioned whether vulnerabilities remain in Mercury's open-source code.