Mike Belshe challenged Claude AI by placing 100 bitcoins in his public wallet
Anthropic disclosed three cybersecurity incidents in a controlled evaluation. The Bitcoin wallet has not been touched so far, and no transfers have been recorded on the chain.
BitGo CEO challenges AI
BitGo CEO Mike Belshe deposited 100 bitcoins in a public wallet and challenged Anthropic's Claude AI model to transfer the funds-after the company disclosed recent cybersecurity incidents. Chain data shows that as of August 2, the wallet has not been touched. Belshe's challenge stems from Anthropic's disclosure that Claude accessed the Internet during a third-party evaluation and hacked into the systems of three real organizations.
Anthropic details three security incidents
Anthropic said the incidents were discovered during a joint review with assessment partner Irregular. According to the company, Claude accessed the Internet in three independent evaluation environments and entered three organizations without authorization. The company said the activities originated from a third-party evaluation environment that was still connected to the Internet. Anthropic added that these models encountered real-life systems while completing specified cybersecurity exercises. The disclosure involved Claude Opus 4.7, Mythos 5 and an internal research model. Anthropic also said it had suspended cybersecurity assessments on July 23 and notified affected organizations on July 27.
Belshe tests AI with public Bitcoin wallet
Following the disclosure, Belshe injected a full 100 bitcoins into a public bitcoin address on July 31. He then challenged Claude to transfer the tokens, saying that if the model successfully completed the transfer, it could keep the tokens. However, blockchain records show that as of August 2, no transfer-out transactions occurred in the wallet. The public address has not changed after the challenge content was spread through social media. This challenge differs from the assessment environment reported by Anthropic: Belshe does not provide system access, does not assign cybersecurity tasks, and does not specify a specific Claude model for testing.
Wallet balance not moved, discussion continues
Anthropic indicates that the reported event involves the model completing the specified Capture Flag exercise rather than attempting to act independently. The company also stated that Claude did not attempt to escape from his environment or replicate himself. Meanwhile, Belshe's public wallet remains open for anyone to monitor via the Bitcoin blockchain. Based on available information, Anthropic has not publicly responded to this challenge as of the time of publication of this article. The incident, which followed OpenAI's disclosure in July of another security review involving Hugging Face, sparked similar discussions about AI cybersecurity.

Exchange Ranking
Top Exchanges
24h Volume Ranking
Popularity Ranking
Exchange BTC Balance
Proof of Reserves
Decentralized Exchanges
Funding Rate
Funding Heatmap
Liquidation Data
Max Pain
Long/Short Ratio
Whale L/S Ratio
Binance/Okex/Huobi L/S
Bitfinex Margin L/S
ETF Tracker
Solana ETF
XRP ETF
Hong Kong ETF
Bitcoin Treasuries
Crypto Reversal
Ethereum Reserves
HyperLiquid Wallet Analysis
Hyperliquid Whale Watch
Large Transactions
On-chain Movement
Bitcoin ROI
Stablecoin Market Cap
Options Analysis
News
Articles
Economic Calendar
Features
Wallet
Contract Calculator
Security
Collections
Watchlist
Following
BTC