Two Ethereum researchers are urging crypto holders to consider a slow, orderly migration of funds to fresh wallet addresses, warning that rapid advances in artificial intelligence could undermine the cryptography securing today’s blockchains well before quantum computers arrive. The recommendation frames the next phase of crypto security as an AI-math race, not a quantum one, and explains exactly how everyday holders can shrink their exposure while the industry works on longer-term fixes.
The call to action centers on the elliptic curve digital signature algorithm (ECDSA), the scheme that signs every on-chain transaction. Researchers argue that AI’s growing strength in pure mathematics could expose weaknesses in ECDSA far sooner than conventional timelines suggest. The suggestion is not panic, but preparation: set funds in motion toward addresses whose public keys have never been broadcast on-chain, then move the rest after each new transaction. Done carefully, this creates a buffer while the cryptography community reacts.
Why AI math changes the timeline for crypto security
The trigger for the warning was a large release of mathematical findings from OpenAI covering algebra, theoretical computer science and mathematical logic. Public demonstrations like that release matter because they show how quickly frontier models are closing gaps in advanced math. Earlier in 2026, the same lab used a team of 10,000 autonomous AI agents working in parallel to solve the Navier-Stokes equation, one of the most famous open problems in mathematics and physics, in just 88 hours. Sudden, tangible progress on hard math problems changes what cryptographers can rely on.
Elliptic curves are a particular concern because they carry rich mathematical structure that can be attacked with specialized techniques, including Schoof’s algorithm, Frobenius-based methods, and pairing-based approaches. By contrast, hash functions are designed to minimize algebraic structure, which is why hashes are widely expected to hold up longer against AI-assisted attacks. This structural difference is what makes wallet signatures, which depend on elliptic curves, the more urgent focus.
Beyond elliptic curves: lattice cryptography is also in scope
The same concern extends to lattice-based cryptography, the family of schemes often promoted as the leading defense against future quantum attacks. A widely held mental model is that elliptic curves will break first, while hashes and lattices remain safe. That model may be too optimistic if AI-driven mathematical progress keeps accelerating.
Concrete lattice-based security could take serious hits from the next two years of AI math, which is one reason Ethereum’s lean roadmap has moved toward a hash-only direction, reducing surface area wherever possible. The implication for projects building on lattice primitives is that “quantum-resistant” no longer means “AI-resistant,” and assumptions on both timelines need to be revisited.
What ‘bunker mode’ looks like for an everyday holder
The practical advice is straightforward and low-cost for most users:
- Generate fresh wallet addresses whose public keys have never been used to sign a transaction. On-chain history is what exposes a key in the first place, so an unused address keeps that exposure minimal.
- Migrate funds gradually, not in one move. A rushed migration can introduce its own risks, including address confusion and operational mistakes.
- Move the remaining balance to a new address after every transaction you sign. Each spend reveals the public key, so rotating afterward keeps the window small.
- Large holders should move first because they hold the biggest targets and have the operational capacity to execute carefully, while retail users with smaller balances and less migration experience follow once the process is proven safe.
- Skip the move if it is not easy to do safely. Botched migrations have historically cost users more than hacks have, so the threshold for action is “simple for me to execute correctly.”
The “bunker mode” framing means a quiet reduction of where risk is sitting while the cryptographic picture becomes clearer.
Why this matters before any quantum threat shows up
The risk being flagged is not quantum at all but conventional mathematics overturning unproven cryptographic hardness assumptions, accelerated by AI tools that outpace human cryptographers’ ability to audit. A “very sober call” from outside the Ethereum research community places that concern squarely within mainstream serious crypto investment thinking rather than on the fringe.
For wallet users, the gain is concrete: a simple habit, applied now, cuts the attack surface of their holdings without waiting for protocol upgrades, new signature schemes, or any single team to ship a fix. For builders, the gain is a clearer research agenda: reduce reliance on elliptic curve and lattice assumptions, lean on hash-based designs, and treat “quantum-safe” and “AI-safe” as two separate bars that both need to clear.
FAQ
What is crypto ‘bunker mode’ and why is it being recommended now?
Bunker mode is the term Ethereum researchers are using for a controlled, gradual migration of crypto funds into fresh wallet addresses whose public keys have not yet been revealed on-chain. It is being recommended now because recent AI progress in mathematics, including a large release of new mathematical findings and a parallel-agent solution to the Navier-Stokes equation in 88 hours, suggests the cryptography behind today’s wallets could face AI-driven attacks sooner than expected.
Which cryptography is most at risk from AI, elliptic curves or lattices?
Elliptic curves, used in ECDSA wallet signatures, are seen as the most exposed because they have rich mathematical structure that specialized algorithms can attack. Lattice-based cryptography, often promoted as quantum-resistant, is also being flagged as potentially vulnerable to AI-accelerated math over the next two years, which is why some roadmaps are moving toward hash-only designs.
Should crypto holders move their funds to new wallets right away?
No one is calling for a rushed move. The recommendation is to set in motion a slow, controlled migration to fresh addresses, start with the largest holders, and rotate addresses after every transaction. Anyone for whom the move is not simple to execute correctly is advised to wait, since botched migrations can cost more than the attacks they are meant to avoid.
See our editorial disclaimer for how our articles are produced.
Run a free scan to see your AI Visibility Score, SEO rating, and local citation accuracy.