Ethereum co-founder Vitalik Buterin has rejected a prediction that artificial intelligence could weaken confidence in Bitcoin’s security and cause its price to fall by more than 50% within the next two years.
Investor and AI-risk commentator Liron Shapira assigned a 50% probability to his forecast that Bitcoin could suffer such a decline as increasingly capable AI systems undermine perceptions of the network’s security and robustness. Buterin took the opposite view, arguing that a direct failure of Bitcoin’s hashing or proof-of-work mechanisms remains extremely unlikely over that period.
Buterin’s response focused primarily on Bitcoin’s technical resilience rather than attempting to forecast its market price. He argued that many AI-related threats would affect the network’s surrounding infrastructure rather than its underlying consensus mechanism and could be addressed through software upgrades without requiring changes to Bitcoin’s broader social consensus.
Bitcoin’s ecosystem includes full-node software, mining pools, individual miners and the internet infrastructure connecting them. AI could potentially make it easier for attackers to discover vulnerabilities in node software, wallet applications, mining systems or communications infrastructure. At the same time, defenders could use similar technology to review code, test patches and identify suspicious activity.
Buterin argued that client developers and mining pools could respond to many such threats through conventional upgrades. He described the likelihood of an “actual” break in Bitcoin’s hashing or proof-of-work system as extremely small, while acknowledging that this assessment does not eliminate other cybersecurity risks affecting users and infrastructure.
Bitcoin’s proof-of-work system relies on SHA-256 to secure the mining process. A fundamental cryptographic attack would require a practical weakness that significantly reduces the computational work needed to produce valid hashes. Current AI systems have not demonstrated such a capability.
The more immediate concern may instead involve infrastructure surrounding the Bitcoin network. More capable AI systems could automate phishing, credential theft, malware development and vulnerability discovery, potentially increasing risks for exchanges, wallet providers, mining pools and node operators.
Recent security incidents illustrate the distinction. A vulnerability affecting some COLDCARD-generated wallets involved insufficient entropy during key generation and exposed affected funds to theft. Such a problem can result in significant losses without compromising Bitcoin’s hashing algorithm or proof-of-work consensus.
The debate also reflects concerns that AI could increase the speed at which attackers discover vulnerabilities. Because Bitcoin’s codebase is publicly available, security researchers and potential attackers can examine the same software. A sufficiently capable AI system could potentially analyze large repositories and test attack paths at a scale that would be difficult for human researchers to match.
Bitcoin security organizations have therefore called for defenders to receive access to advanced AI tools. In August, the Bitcoin Policy Institute and more than 40 digital-asset organizations urged leading AI laboratories to provide vetted open-source security teams with controlled access to frontier models, computing resources and secure testing environments.
The initiative was supported by companies and organizations including Block, Coinbase, Strategy, MARA, Galaxy, BitGo, Brink and Trezor, as well as several Bitcoin development groups. Supporters argued that security researchers should have access to comparable tools before more capable AI systems become widely available to potential attackers.
The underlying concern is less about AI directly breaking Bitcoin’s proof of work and more about whether defenders can identify and fix weaknesses quickly enough as attackers gain more sophisticated automated capabilities.
Buterin also distinguished between operational vulnerabilities and deeper failures that could require broad network coordination. Conventional problems involving node software or mining infrastructure can generally be addressed through upgrades. A fundamental failure of Bitcoin’s cryptography would be considerably more complicated.
If attackers were able to forge digital signatures or bypass proof-of-work requirements, Bitcoin developers could potentially need to introduce new cryptographic standards and coordinate a network-wide migration. Nodes, miners, exchanges, custodians and users would then have to agree on which software and transaction history to recognize, creating the possibility of significant disruption or a network split.
Quantum computing represents another potential long-term threat to Bitcoin’s cryptographic security, although it is fundamentally different from AI because it relies on a different computing architecture rather than machine-learning systems. Developers have already explored potential quantum-resistant approaches, but the timing and severity of any practical quantum threat remain uncertain.
For now, there is no evidence that AI has compromised Bitcoin’s underlying security mechanisms. Shapira’s 50% price-crash forecast is a probabilistic prediction rather than a technical finding, and it does not identify a specific AI capability or attack sequence that would necessarily produce a 50% decline.
Likewise, Buterin’s assessment is not a formal risk model or guarantee. His argument is that Bitcoin’s ecosystem can adapt to many conventional security threats through software upgrades, while a direct compromise of its hashing or proof-of-work foundations remains highly unlikely.
Bitcoin traded near $79,590 on Sept. 7, down approximately 0.5% during the latest session. There is no evidence that the move was directly related to the debate, as Bitcoin prices are influenced by factors including liquidity, derivatives positioning, ETF flows, macroeconomic conditions and broader market sentiment.
The disagreement ultimately centers on how quickly AI could change the balance between attackers and defenders, rather than evidence that artificial intelligence has already created a fundamental threat to Bitcoin’s consensus system.
