Former pharmaceutical executive and convicted securities fraudster Martin Shkreli has weighed in on the quantum computing debate, asserting that Shor's algorithm—not artificial intelligence—is the most credible path to cracking Bitcoin's elliptic curve cryptography, provided quantum hardware scales sufficiently. However, he emphasized that today's machines are far from capable.
Separating Reality from Hype
During an appearance on the Bitcoin Rails podcast #38 hosted by Isabel Foxen Duke, Shkreli drew a clear line between quantum computing's promise and its current limitations. He stated that quantum will not replace classical computing (e.g., Nvidia GPUs), but “when it comes to Shor's algorithm and Bitcoin, you have something to worry about.” The catch: practical quantum machines remain slow, noisy, and fragile, with error rates that must improve by orders of magnitude for any credible attack.
The core issue is fidelity. Each quantum logic gate succeeds with some probability; best-in-class gates achieve about 99.99% fidelity. While that sounds impressive, compounding error across the millions of gates required for a full Shor run causes accuracy to collapse. The path forward involves either building vastly cleaner physical qubits or layering powerful error-correction codes to create reliable “logical” qubits—both daunting challenges.
Shkreli cited IBM's publicly accessible ~150-qubit systems as a reality check: they are educational tools, nowhere near the logical-qubit counts needed for a Bitcoin-scale break. He estimated that targeting Bitcoin's 256-bit elliptic curve would require roughly a million logical qubits, implying hundreds of millions to a billion physical qubits given error-correction overhead. Today's machines are orders of magnitude smaller.
From Prison to DeFi
Since his release from prison in 2022, Shkreli has immersed himself in cryptocurrencies, focusing on decentralized finance (DeFi) and blockchain mechanics. He revealed in an X Spaces chat that he used Uniswap—Ethereum's leading decentralized exchange—while incarcerated, captivated by its ability to bypass traditional financial gatekeepers. He expressed bullish views on Ethereum, Solana, and Algorand, predicting that Ethereum could eventually flip Bitcoin in market capitalization.
Technical Gaps and Timelines
Shkreli noted that even with improved physical qubits (mentioning fluxonium as a candidate), quantum teams must add many more “nines” of reliability. He contrasted classical GPUs, which operate with extremely low error rates, with quantum hardware still wrestling with noise, decoherence, and even cosmic-ray-induced bit flips. Until error correction tames these issues, Shor's algorithm remains a whiteboard triumph.
He also stressed a crucial nuance: quantum computers are not fast in clock speed (often measured in kilohertz or less), but they are valuable because algorithms like Shor's change the complexity from exponential to polynomial time. That complexity drop is the point; the hardware simply isn't there yet.
On timelines, Shkreli avoided concrete predictions, acknowledging that a credible Shor-class attack on Bitcoin is not a five-year story and could take decades, given the gap between today's physical qubits and tomorrow's error-corrected logical qubits.
He conceded that non-quantum routes—mathematical breakthroughs potentially aided by AI—cannot be ruled out, but he still ranks quantum as the likelier first mover against elliptic curve cryptography. Either way, he framed the target as the cryptography itself, not “putting Nvidia out of business.”
Ethics: Hacking Satoshi's Coins?
When asked about the hypothetical of hacking Satoshi Nakamoto's coins, Shkreli said the intellectual achievement is the draw; he “wouldn't want to hold those tokens,” calling that theft even if mathematically successful. He suggested that research findings could be published without looting anyone's wallet.
FAQ
- What exactly does Shor's algorithm threaten? It targets the hard math (factoring/discrete logs) behind elliptic curve cryptography used by Bitcoin wallets, assuming a large, error-corrected quantum computer exists.
- Why aren't today's quantum machines a risk? Error rates compound across millions of gates, and current fidelities fall short for a full Shor run.
- How many qubits would an attack need? Shkreli discussed roughly a million logical qubits, implying hundreds of millions to a billion physical qubits with error correction.
- Could AI break Bitcoin's crypto first? Shkreli says advances in mathematics aided by AI are possible, but he still views quantum as the likelier first mover against ECC.