AI Threat to Crypto Security: Mythos Model Sparks Concerns, But Bitcoin Investors Remain Unfazed

Artificial intelligence is increasingly seen as a looming threat to cryptocurrency security, yet institutional investors are not backing down. According to Coinbase, the new AI model could autonomously exploit security flaws, potentially destabilizing crypto markets.

Anthropic’s Claude Mythos AI model, introduced in early 2025, has the capability to identify and exploit zero-day vulnerabilities—undisclosed security flaws in operating systems and web browsers. This raises concerns about its potential misuse in the crypto sector, which relies heavily on browsers, wallets, and open-source tools for transactions.

Coinbase Stands Firm on Bitcoin Despite AI Threats

David Duong, Coinbase’s global head of investment research, emphasized that institutional confidence in Bitcoin remains strong. In a recent note, Duong stated that 75% of institutions still view Bitcoin as undervalued, despite short-term market volatility.

Duong’s optimism contrasts with Bitcoin’s recent performance. The cryptocurrency is still down nearly 40% from its October 2024 peak, while the S&P 500 reached record highs in April 2025. Bitcoin is currently trading at $77,000, following a 13% rally over the past month.

“Our outlook for the next quarter remains neutral, but there are early signs the market may be finding a floor,” Duong said. “Much of the speculative activity has been flushed out.”

ETF Inflows Signal Renewed Institutional Interest

Exchange-traded fund (ETF) data supports Duong’s assessment. April 2025 marked the best month for Bitcoin ETFs since October 2024, with nearly $2 billion flowing into these funds over the past 30 days, according to DefiLlama.

How Dangerous Is Mythos? Crypto Exchanges Rush to Adopt AI

The urgency around Mythos stems from its potential to disrupt crypto security. Major exchanges, including Coinbase and Binance, are reportedly racing to access the model, despite its limited release to a small group of tech partners due to misuse concerns.

Anthropic has not set a public release date for Mythos, citing its powerful capabilities. The model’s ability to find and exploit zero-day vulnerabilities could pose a severe risk to crypto infrastructure, which is already a prime target for cyberattacks.

Cybersecurity Experts Warn of Rising Risks

Deddy Lavid, head of cybersecurity firm Cyvers Alert, highlighted the seriousness of the threat. “If AI can find weaknesses across the internet at scale, crypto could be one of the first sectors hit,” he told DL News in April 2025. The industry’s reliance on browsers, wallets, and open-source tools makes it particularly vulnerable.

The stakes are high. The crypto industry has already suffered losses of approximately $3.4 billion to hacks and attacks in 2025, according to Chainalysis.