Weekly Review: Could AI Crack Bitcoin's Signatures Before Quantum Computers?
Key Takeaways
- •Ethereum Foundation researcher Justin Drake argues AI could break the ECDSA signatures used by Bitcoin and Ethereum before quantum computers do, though no evidence of such a breakthrough currently exists.
- •The privacy coin market expanded from USD 6.2 billion to roughly USD 30 billion in one year, while the rest of the crypto market excluding stablecoins lost about 35%.
- •The ECB outlined three models for tokenized central bank money and has already linked 13 financial institutions, including Deutsche Bank and Santander, to settle tokenized securities trades directly in central bank money.
- •FinCEN withdrew its never-enforced proposals on unhosted wallets and crypto mixers, attributing the move to the Trump administration's deregulation agenda, but it still views mixers as a money laundering risk and has not ruled out further action.
- •Automated market makers on decentralized exchanges allow liquidity providers to earn trading fees through self-rebalancing pools, but volatility-driven impermanent loss means liquidity provision is not a risk-free income source.

Artificial intelligence taking aim at cryptographic signatures, a fast-growing privacy coin sector, a new ECB blueprint for tokenized central bank money, and a regulatory rollback in Washington: these were the defining developments of the past week in blockchain and cryptocurrencies.
Ethereum Researcher Warns AI Could Break Crypto Encryption
Ethereum Foundation researcher Justin Drake believes artificial intelligence could break the cryptographic signatures securing crypto before quantum computers manage to do so. The scheme in question is ECDSA, the Elliptic Curve Digital Signature Algorithm that Bitcoin and Ethereum use to sign transactions. These signatures are the mathematical proof that the holder of a private key authorized a transaction; a working break would let attackers forge such authorization. Quantum computers have long been viewed as the eventual threat here, since algorithms such as Shor's could derive private keys from exposed public keys—Drake's argument is that AI could get there first.
The trigger for his warning is a release by OpenAI: 722 AI-generated mathematics manuscripts that nobody has independently verified yet. In Drake's assessment, elliptic curves are especially vulnerable because of their algebraic structure. In the worst case, ECDSA could therefore fall within months rather than years. However, there is no evidence for such a breakthrough.
As a precaution, Drake recommends what he calls Bunker Mode. Because the public key only becomes visible during signing—addresses conceal it behind a hash—remaining funds should be moved to a fresh address after every signature. In addition, companies such as Binance, Robinhood, and Tether should harden their cold storage. Nevertheless, he advises against haste, since a rushed migration would do more harm than good. In the long run, Ethereum aims to move fully to hash-based cryptography, an approach whose security rests on hash functions alone and is considered resistant to quantum attacks.
Privacy Coins: The Missing Layer Between Crypto and Institutions
On public blockchains, every balance and every counterparty remains permanently visible. This is an obstacle for institutions, because no fund announces its trades publicly—and anyone on-chain can see and front-run such positions. Privacy therefore counts as the missing layer between crypto and institutions, and investors are already pricing it in: the privacy coin market grew from USD 6.2 to around USD 30 billion in one year. By contrast, the rest of the crypto market, excluding stablecoins, lost around 35%.
The Canton Network shows one possible solution. There, positions remain hidden from competitors, although supervisors can access them when needed. Goldman Sachs, J.P. Morgan, and the US custodian DTCC already use it. Zcash, the largest privacy network, accounts for around USD 20 billion of the total.
Competition, however, is growing. Ethereum puts privacy at the center of its 2026 roadmap, and Solana encrypts amounts with Confidential Balances. This could erode the premium of dedicated privacy coins like Zcash.
ECB Sees Three Paths to Digital Central Bank Money (CBDC)
Isabel Schnabel of the European Central Bank (ECB) Executive Board presented three models for blockchain-based central bank money in London. In the first, the central bank tokenizes bank reserves directly. The second keeps the existing payment system in place and links it to the blockchain via an interface. In the third, a private intermediary issues tokens fully backed by central bank reserves, although for holders they remain a private claim. All three models pursue the same goal: keeping central bank money as the settlement anchor in tokenized markets. That matters because settlement in central bank money carries no credit risk between banks, which is why policymakers treat it as the foundation for tokenized financial markets.
In practice, the ECB has gone further. It recently began linking blockchain platforms to its payment system, allowing banks to settle tokenized securities trades directly in central bank money. At launch, 13 financial institutions joined, including Deutsche Bank and Santander. Moreover, the ECB plans to invest a small part of its own funds in such securities, although it did not name an amount.
FinCEN Abandons Its Planned Crypto Travel Rule
The US anti-money laundering agency FinCEN has withdrawn two crypto proposals that never took effect. The first dates from December 2020 and concerned self-custody wallets, known as unhosted wallets: financial institutions would have had to record counterparty data from USD 3,000 upward. The second proposal, from 2023, classified crypto mixing as a primary money laundering concern. Mixers blend the transactions of many users to obscure payment trails, and the planned reports would also have included users' IP addresses.
FinCEN attributes the withdrawal to the Trump administration's push to cut regulation. However, this is no all-clear: the agency still sees mixers as a money laundering risk and has not ruled out further action. And because neither proposal ever took effect, the withdrawal leaves institutions' current US obligations unchanged.
Nothing changes for Swiss institutions in any case. FINMA already requires proof that the customer controls a self-hosted wallet, regardless of the amount. Meanwhile, a similar requirement applies in the EU from EUR 1,000.
Automated Market Makers as Decentralized Portfolio Managers
Also worth knowing: on decentralized exchanges, users trade against a token pool rather than against each other. A fixed formula, not an order book, determines the price. This mechanism is called an automated market maker (AMM), a model popularized by decentralized exchanges such as Uniswap. The capital in the pool comes from investors known as liquidity providers, who in return receive a share of the trading fees.
Anyone who provides liquidity therefore holds a portfolio that rebalances itself with every trade. If a token rises, the pool sells part of it; if it falls, the pool buys more. At its core, an AMM thus works like an asset manager with a fixed investment rule.
However, this comes at a price. When a token rises sharply, simply holding it would often have earned more. This difference, known as impermanent loss, grows with price volatility. Admittedly, fees can offset the loss. Nevertheless, providing liquidity is not a risk-free source of income.
This weekly review was originally published on Crypto Valley Journal.