Haseeb Qureshi, a managing partner at the crypto venture capital firm Dragonfly, has predicted that annual losses from decentralized finance (DeFi) hacks will decrease in 2026 compared to the previous year, even as concerns about artificial intelligence (AI)-powered attacks persist. His assessment provides a counterpoint to widespread fears that AI would trigger a surge in sophisticated crypto theft.
Why Losses Are Falling Despite More Hacks
According to Qureshi, the number of hacking incidents targeting DeFi protocols has actually increased. However, the average financial loss per incident has been steadily declining. This trend suggests that attackers are increasingly focusing on smaller, less secure, or defunct protocols rather than the major, well-fortified platforms that hold the bulk of user funds.
Qureshi noted that losses stemming from compromised admin keys and multisignature wallets have also decreased. This indicates that the industry’s operational security practices are maturing. Large protocols, in particular, have invested heavily in hardening their systems against both conventional exploits and emerging AI-based attack vectors.
The Unrealized AI Threat in DeFi
While the potential for AI to automate vulnerability discovery or execute complex, adaptive attacks has been a topic of intense discussion, Qureshi stated that these fears have not yet materialized into a major crisis. The lack of a catastrophic AI-driven hack in 2025 supports the view that current AI models are not yet capable of reliably breaking the defenses of top-tier DeFi protocols at scale.
This is not to say the threat is nonexistent. Security teams are actively developing AI-driven defensive tools to monitor for anomalous behavior and patch vulnerabilities faster than human teams alone could manage. The ongoing arms race between attackers and defenders is shifting toward automation on both sides.
What This Means for DeFi Users and Investors
The declining average loss per hack is a positive signal for the ecosystem’s overall health. It implies that the most valuable assets are becoming safer, and that security best practices are becoming standard among serious projects. For users, this reinforces the importance of sticking with established, audited protocols rather than chasing high yields on unaudited or anonymous platforms.
For investors and developers, the data suggests that while the number of attacks may not drop, the financial impact is being contained. This could reduce the risk premium associated with DeFi investments, potentially attracting more institutional capital into the space.
Conclusion
The prediction from Dragonfly’s Haseeb Qureshi offers a cautiously optimistic outlook for DeFi security in 2026. While the threat landscape remains active, the industry’s response—through improved defenses and a shift in attacker focus toward smaller targets—appears to be effectively limiting overall financial damage. The anticipated AI-driven crisis has not arrived, giving the sector more time to build robust, automated security systems for the future.
FAQs
Q1: Why are DeFi hack losses expected to fall in 2026?
Because the average loss per hacking incident is declining, even though the total number of incidents is rising. Hackers are targeting smaller, less secure protocols, while major platforms have strengthened their defenses.
Q2: Has AI made DeFi hacking worse?
Not yet. Despite widespread fears, a major AI-driven hacking crisis has not materialized. Current AI models are not reliably breaking the defenses of top-tier DeFi protocols at scale.
Q3: What can DeFi users do to stay safe in 2026?
Users should prioritize established, audited protocols over anonymous or unaudited platforms that offer high yields, as these are the primary targets for current attacks.