As concerns grow that advances in artificial intelligence could undermine core cryptographic assumptions used in crypto wallets, Dragonfly managing partner Haseeb Qureshi has challenged the idea that users can simply “bunker” funds by moving them to new addresses. In a Thursday X post, Qureshi described “bunker mode” as “cryptographic doomerism” and argued that address migration alone would not solve the systemic risk if widely used signature schemes are compromised.
Qureshi also urged networks to prepare for a scenario where cryptographic signatures can be broken—proposing a “Cryptographic Recovery Mode” that would rely on hash-based backups and allow validators to execute recovery if signatures fail. The comments followed warnings from Ethereum researcher Justin Drake that ECDSA—an elliptic-curve signature algorithm widely used across crypto—could be compromised “in the worst case, in months, not years,” citing an OpenAI report on progress in mathematics.
Key takeaways
- Dragonfly’s Haseeb Qureshi says “bunker mode” (moving funds to new addresses) doesn’t remove risk if other users’ coins are also compromised and dumped.
- Qureshi proposes “Cryptographic Recovery Mode,” using hash-based backup signatures that validators could use to force recovery if signatures are broken.
- Glassnode data cited by Qureshi indicates millions of bitcoins are potentially exposed—either due to address reuse or specific address formats.
- Ethereum researcher Justin Drake warns ECDSA may fail sooner than expected, framing the timeline as potentially “months, not years.”
Why “bunker mode” may not be enough
Qureshi’s core objection is that address-level protection only works if a broken cryptographic system can’t be used to impact the broader market. In his Thursday post, he wrote that “bunker mode” would protect investors’ coins only as long as those assets are never moved away from the “fresh address” where exposure is reduced.
He added that the approach becomes economically fragile if other coins are also at risk. In Qureshi’s framing, even if one user avoids immediate compromise, the value of their protected holdings could deteriorate in a scenario where “all of the other coins are being hacked and mass-sold,” leaving those protected coins “worthless.”
Rather than focusing on individual hygiene—where users might rotate addresses to reduce public-key exposure—Qureshi argued for protocol-level resiliency. That shift matters because it acknowledges that a cryptographic failure could be systemic: if the signature mechanism underpinning wallets and transactions becomes unreliable, the threat isn’t confined to one address pattern or one user’s behavior.
Glassnode: millions of bitcoins may be exposed
The debate is occurring alongside data suggesting a large portion of Bitcoin’s supply could be vulnerable to public-key or address-pattern risks. Qureshi referenced Glassnode’s analysis that 6.26 million BTC are in “vulnerable addresses,” which he presented as a potential consequence of how keys and address information are reused or encoded.
According to Glassnode, 4.33 million BTC are exposed due to address reuse—moving those coins to a fresh address would, in theory, end that specific form of exposure. An additional 1.94 million BTC are exposed through their address format, according to Glassnode co-founder Rafael Schultze-Kraft, who shared further breakdown details in a Thursday X post.
The exposure picture is not limited to cold storage. Qureshi noted that nearly 1.8 million BTC from the exposed supply are held on cryptocurrency exchanges. He cited Glassnode’s observation that 57% of all exchange balances are currently exposed, underscoring that operational custodianship may face higher immediate pressure if cryptographic assumptions fail.
Drake’s warning: ECDSA could break sooner than expected
These comments build on earlier warnings from Ethereum researcher Justin Drake. In a Wednesday X post, Drake said it was reasonable to “brace” for the possibility that ECDSA could break before the hypothetical “qday” event—when quantum capabilities would be strong enough to threaten modern cryptography. Drake’s timeline, in the “worst case,” was “in months, not years.”
Drake linked his concerns to a Tuesday OpenAI report about the pace of AI progress in mathematics. He also advised users to gradually move funds to fresh wallets where their public key is not exposed, arguing that AI advancements might enable cryptanalytic breakthroughs before quantum computers arrive.
Ethereum co-founder Vitalik Buterin agreed with the need to take the “AI-accelerated math” risk seriously, while he did not recommend users rush to move funds immediately to new wallets. That difference highlights a tension in the community’s approach: whether to optimize for short-term exposure reduction at the user level, or to prioritize longer-term protocol resilience that assumes widespread, coordinated risk.
Qureshi’s “Cryptographic Recovery Mode” proposal
Qureshi’s answer to “bunker mode” is a structured recovery pathway for the moment cryptographic signatures can no longer be relied upon. He proposed a “Cryptographic Recovery Mode,” described as a hash-based backup signature plan that users could map to their addresses.
In this concept, validators would have the ability to force recovery if signatures are broken. The practical importance is that recovery would be designed for a failure state—shifting the focus from trying to prevent every exposure to ensuring the system can continue operating when cryptographic mechanisms do not hold.
While address rotation reduces certain kinds of exposure, Qureshi’s framing implies that such steps may not protect against an environment where attackers can mass-exploit compromised signature functionality at scale. For investors, this translates into a key question: if a cryptographic failure becomes widespread, will value preservation depend primarily on individual wallet practices—or will it depend on how quickly and effectively networks can introduce recovery or validation mechanisms?
Readers watching this space should focus on whether the industry’s response moves beyond guidance for users and toward implementable protocol designs. The next signal to watch is whether researchers and core developers converge on concrete recovery mechanisms—especially ones that can be adopted without requiring every user to act perfectly and immediately under uncertainty.
Be the first to comment