Crypto Firms Unite to Access Cutting-Edge AI Models Amid Rising Security Threats

The Bitcoin Policy Institute (BPI) has coordinated an effort among over 40 crypto companies to request access to cutting-edge AI models from top labs, including OpenAI and Anthropic. These models are currently only available to hackers, who use them to launch sophisticated attacks on Bitcoin's infrastructure.

The BPI argues that allowing approved security experts to use these models would give defenders a new edge in the battle against attackers. However, the letter acknowledges that these models have dual-use capabilities, meaning they can be used for both defensive and malicious purposes.

A report by SlowMist found that there were 182 security incidents on blockchain networks in the first half of 2026, resulting in losses of around $956 million. This is an increase from 121 incidents and $2.373 billion in losses during the same period in 2025.

The BPI's request is part of a growing concern about the use of AI in cryptocurrency security. The U.S.-China AI race is also playing out, with Chinese models lagging behind American ones but still posing a threat to Bitcoin's safety.