A groundbreaking research paper published on February 15, 2024 by CoinMetrics head of research and development Lucas Nuzzi sends a powerful message to anyone questioning Bitcoin’s security: mounting a 51% attack on the Bitcoin network would cost at least $20 billion in hardware alone, and the attacker would still walk away with zero profit. The study, titled “Breaking BFT,” introduces a novel analytical framework that quantifies the true cost of compromising the world’s most valuable blockchain network.
TL;DR
- CoinMetrics publishes “Breaking BFT” study by Lucas Nuzzi, introducing the Total Cost to Attack (TCA) framework
- A 51% attack on Bitcoin would require approximately $20 billion in ASIC hardware procurement
- An Ethereum attack would cost over $34 billion and take six months to execute
- Researchers find no profitable attack strategy for either Bitcoin or Ethereum
- The study provides the first empirical evidence of a Nash Equilibrium in blockchain security
The Total Cost to Attack Framework
Nuzzi and his team at CoinMetrics develop what they call the Total Cost to Attack, or TCA, a comprehensive methodology that goes far beyond simplistic calculations of mining power. The TCA framework accounts for the full spectrum of expenses an attacker would face when attempting to seize majority control of a blockchain network, including hardware procurement, operational costs, electricity, and logistics.
The study leverages data from MINE-MATCH, a comprehensive database of mining hardware, combined with historical ASIC market trends to build realistic attack scenarios. This approach provides a far more accurate picture than previous estimates that focused solely on hash rate rental costs.
Bitcoin’s $20 Billion Shield
For Bitcoin, the research reveals that the cost of acquiring the necessary ASIC miners to execute a 51% attack could surge to $20 billion — and that figure represents only the hardware procurement expense. The study emphasizes a critical point: these ASICs are not readily available for purchase in the quantities required. Nuzzi explicitly states that the supply constraint alone makes the attack impractical.
But what about a nation-state with manufacturing capabilities? The researchers explore this scenario as well, concluding that even if a government attempted to manufacture ASICs from scratch, the only mining hardware model that could be plausibly reverse-engineered is the Antminer S9 — and the manufacturing cost for sufficient units would still exceed $20 billion. This finding underscores not just the financial barriers but also the significant technical and logistical hurdles involved.
Ethereum’s $34 Billion Defense
The study also examines Ethereum’s proof-of-stake security model following the Merge. Attacking Ethereum presents an entirely different set of challenges compared to Bitcoin. The researchers estimate the cost at over $34 billion, factoring in the need to operate more than 200 nodes and spend approximately $1 million on Amazon Web Services infrastructure alone.
One of the study’s key findings challenges a common assumption in the crypto community: liquid staking derivatives cannot be leveraged to gain access to block templates. This means an attacker cannot shortcut the process by accumulating staked ETH through DeFi protocols.
Furthermore, the research estimates that a full attack on Ethereum would require approximately six months to execute due to the network’s churn limit, which prevents large amounts of stake from being deployed all at once. This time delay serves as an additional deterrent, as it gives the community ample opportunity to respond before the attack reaches critical mass.
No Profit, No Attack
Perhaps the most significant finding from the CoinMetrics study is the conclusion that no profitable attack strategy exists for either Bitcoin or Ethereum. The researchers explore various attack scenarios including double-spend attacks and maximal extractable value exploits, finding that none of these strategies would generate enough revenue to justify the enormous upfront investment.
This economic disincentive represents the core of Bitcoin and Ethereum’s security model. The networks do not rely solely on the technical difficulty of mounting an attack but on the fundamental economic irrationality of doing so. When the cost of attacking a network far exceeds any potential gain, rational actors are naturally deterred.
Nash Equilibrium in Blockchain Security
The study identifies what Nuzzi describes as the first empirical evidence of a Nash Equilibrium in the security dynamics of Bitcoin and Ethereum. In game theory, a Nash Equilibrium occurs when no player can improve their outcome by unilaterally changing their strategy. In the context of blockchain security, this means that adversarial actions become unattractive when compared to honest participation.
This finding carries profound implications for the long-term viability of decentralized networks. It suggests that Bitcoin and Ethereum have reached a level of economic security where attacks are not just technically difficult but economically irrational — a self-reinforcing equilibrium that strengthens as the networks grow.
Context: Bitcoin at $52,000
The study arrives at a time when Bitcoin trades at approximately $52,000, having surged past the $50,000 mark earlier in February 2024. The cryptocurrency’s market capitalization exceeds $1 trillion, and the broader crypto market has reclaimed the $2 trillion total capitalization level. Institutional demand through spot Bitcoin ETFs continues to accelerate, with BlackRock’s IBIT surpassing 100,000 BTC in holdings and total ETF inflows reaching $477 million.
Coinbase Custody holds over 90% of all Bitcoin ETF assets across eight of the eleven spot ETF products, further concentrating institutional Bitcoin holdings within regulated infrastructure. This growing institutional presence adds another layer of legitimacy and demand to the network, reinforcing the economic security that the CoinMetrics study quantifies.
Why This Matters
The “Breaking BFT” study provides the most rigorous economic analysis to date of what it would actually take to compromise Bitcoin or Ethereum. The $20 billion and $34 billion price tags, combined with the impossibility of profiting from such attacks, deliver a powerful validation of the security assumptions underlying the two largest cryptocurrency networks.
For institutional investors evaluating Bitcoin through the lens of spot ETFs, this research offers concrete data supporting the network’s resilience. For the broader crypto community, the identification of a Nash Equilibrium in blockchain security provides a theoretical foundation for what many have long believed: that Bitcoin and Ethereum have grown too large and too economically secure to be meaningfully threatened by 51% attacks.
Disclaimer: This article is for informational purposes only and does not constitute financial advice. Cryptocurrency investments carry significant risk, and past performance does not guarantee future results. Always conduct your own research before making investment decisions.
$20B in ASIC hardware alone and still no profit. the nash equilibrium finding is huge for the security argument
the paper assumes you can even source that many ASICs without the manufacturers noticing. real world attack costs are probably even higher
Nuzzi assuming ASIC sourcing would be noticed is the key insight. you cant secretly buy $20B in hardware without bitmain and friends knowing
Sanja P. bitmain would absolutely notice. their entire order book would flag a $20B request. you cant hide hardware acquisition at that scale
Sanja P. exactly. bitmain and microbt would notice a $20B order before it was 5% fulfilled. the sourcing assumption is conservative
twenty_bill the Nash Equilibrium part is what matters most. attacking BTC isnt just expensive, its economically irrational. the paper proves what maxis always claimed
the nash equilibrium finding is what makes this paper actually matter. its not just expensive to attack BTC, its irrational
eth at $34B attack cost is actually higher than BTC. makes sense given the staking dynamics
ETH attack cost at $34B is wild. proof of stake is actually harder to attack than most people assume
hash_wars_ proof of stake attack costs are higher because you need to acquire ETH on the open market. buying $34B of ETH is not subtle either
Nuzzi included the difficulty adjustment in his model which most attack cost estimates completely ignore. thats why the numbers are so much higher than back of napkin math
Lieselotte H. even with perfect hardware acquisition the chain would fork and your ASICs mine worthless coins. the attack succeeds and fails simultaneously
fork_resist_ the chain fork point doesnt get enough attention. you spend 20B on ASICs, attack succeeds, then the community forks and your hardware mines a dead chain. heads you lose tails you lose
Lieselotte H. difficulty adjustment being included in the model is what sets this apart from every other attack cost estimate. most back of napkin math just multiplies hashprice by 51 percent
the Nash Equilibrium finding is the strongest case for BTC security. attacking bitcoin devalues the coin you just spent $20B to mine. self defeating by design
Nuzzis Nash Equilibrium finding proves what maxis always said. attacking BTC isnt just expensive, its self-defeating by design
even if you somehow sourced $20B in ASICs silently, the difficulty adjustment would crush your margins before you recovered a fraction of the cost. the math just doesnt work for attackers
pool_hashrate_ the difficulty adjustment point is key. even if you sourced the ASICs you would be mining at reduced output for weeks before the adjustment caught up. attack economics get worse every halving
pool_hashrate_ difficulty adjustment alone kills the attack economics. you mine at reduced output for weeks before it catches up. add ASIC depreciation and its pure loss
difficulty adjustment alone makes the attack economics terrible. you mine at reduced output for weeks before it catches up. pure loss on top of 20B capex
rig_logistics_ difficulty adjustment crushing attack margins is the real moat. even with free ASICs the economics dont work
the Nash Equilibrium finding is what makes this paper genuinely important. attacking BTC isnt just expensive, its self defeating. the attacker devalues their own position
the nash equilibrium argument is the strongest case for BTC security. even with 20B in ASICs you devalue your own position. self defeating by design
Henrik J. the nash equilibrium is the strongest argument for BTC security. even if you had 20B in ASICs attacking btc devalues the coin youre mining. self defeating
20B in hardware and bitmain would notice at 1B. supply chain transparency makes a covert attack basically impossible
Stefan V. the Bitmain order book point is critical. you cannot source $20B in ASICs without every manufacturer knowing exactly what you are doing