Why Bitcoin by Satoshi Nakamoto cannot be hacked

Updated:
✦ Ask AI about this article
Why Bitcoin by Satoshi Nakamoto cannot be hacked

On October 29, 2026, about $6.3 billion was withdrawn from Binance's hot wallets within a few dozen minutes — according to the exchange itself, this is the largest theft in cryptocurrency history. Among the stolen assets were over 18,000 BTC. Meanwhile, approximately one million bitcoins attributed to Satoshi Nakamoto have been lying on open addresses on the blockchain for over 17 years. Their balance is visible to everyone. And no one has taken them.

The paradox disappears as soon as we separate two things that are constantly mixed in the news: cryptographic security and the security of the system around the keys. Binance was not hacked due to weak mathematics — the signatures in the stolen transactions were absolutely valid. It's just that the wrong person signed them.

In this article, we will explore what protects Satoshi's bitcoins, why artificial intelligence is powerless here, how real the quantum threat is, and what actually leads to the loss of cryptocurrency. And about how AI is already helping to attack crypto today — not by breaking keys, but through weaknesses around them — in a separate article "How AI Helps Hackers Attack Cryptocurrencies".

Contents

🔓 Binance Lost Billions. Why It's Not a Bitcoin Hack

The attack began on the evening of October 29, 2026 (around 20:05 UTC) and affected Binance's hot wallets on approximately ten blockchains simultaneously: Ethereum, Bitcoin, Tron, Arbitrum, Solana, and others. According to TRM Labs' calculations, 18,270 BTC (about $1.47 billion) were withdrawn from the Bitcoin wallet alone.

What is known from preliminary data:

  • Binance stated that it found no evidence of private key leaks from the wallets. Initial investigation points to a compromised DevOps engineer's laptop, through which the attacker gained access to a service container and initiated fund withdrawals from there.
  • TRM Labs and Elliptic analysts estimate that the attacker forced Binance's own signing system to sign transactions in their favor. In Ethereum, 50 pre-signed EIP-7702 authorizations were used for this, delegating the wallet to the attacker's contract. In other networks, funds were transferred via regular signed transfers.
  • The exchange promised to cover all losses from the SAFU fund and its own funds, and temporarily suspended withdrawals.
  • TRM Labs, with high confidence, links the attack to the North Korean group TraderTraitor, but as of October 30, no government has officially confirmed this attribution.

🔑 The main point: no algorithm was broken. ECDSA worked flawlessly — the system signed exactly what was presented to it for signing. The attack targeted the path to the signature: workstation → service → signer. This is an infrastructural, not a cryptographic vulnerability.

A centralized exchange is thousands of servers, employees, CI/CD pipelines, and internal services that have the right to initiate transfers. Each of them is part of the attack surface. Satoshi's coins have no such surface: no server, no employee, no API. There is only mathematics.

💰 How Many Bitcoins Does Satoshi Have and Where Does This Figure Come From

An important clarification that often gets lost in headlines: there is no single "Satoshi's wallet" as one entity. There are tens of thousands of individual outputs — rewards for mining early blocks in 2009–2010, which researchers attribute to a single miner.

What is the Patoshi pattern

In 2013, researcher Sergio Demian Lerner noticed a characteristic "fingerprint" of a single miner in the early blocks: a specific behavior of the ExtraNonce field and other technical signs that distinguished it from all other network participants. This fingerprint was named the Patoshi pattern. The miner with this pattern started mining from the very launch of the network and, by estimates, found about 22,000 blocks.

The connection "Patoshi = Satoshi" is a strong hypothesis, but not proof. The arguments in its favor are indirect: the miner worked from day one, and known transactions that Satoshi made early on originate from blocks with this pattern.

Estimates vary

SourceEstimateComment
Arkham Intelligence≈ 1,096,362 BTCThe most frequently cited figure (~1.1 million)
Bitquery (September 2026)from ~0.9 to ~1.17 million BTCThe range depends on how strictly the pattern is interpreted
Bitquery, "provably immobile"1,023,352 BTCCoins with the pattern that have not moved as of September 1, 2026

Bitquery rebuilt the fingerprint from scratch, analyzing 54,316 early blocks, and achieved a 99.2% block match with the public Patoshi list. According to their data, 99.86% of these coins have never moved, and only about 1,550 BTC have been spent from the set over time — mostly when Bitcoin was still an experiment.

In dollars, this is over $80 billion at the exchange rate at the end of October 2026 (after the Binance hack, BTC dropped below $77k). The exact amount changes daily with the exchange rate, so it is more correct to provide the number of BTC rather than a dollar valuation in articles.

📌 Honest phrasing: not "Satoshi has 1.1 million BTC," but "about 1.1 million BTC are attributed to a miner who is highly likely identified with Satoshi."

🔐 How Bitcoin Cryptographic Protection Works

Private and Public Key

A Bitcoin private key is simply a random 256-bit number k within the order of the curve n (slightly less than 2256). The public key is a point on the secp256k1 elliptic curve, obtained by multiplication:

P = k · G

where G is a fixed generator point, the same for everyone. Computing P from k takes milliseconds on any laptop. But finding k knowing P and G is the Elliptic Curve Discrete Logarithm Problem (ECDLP), for which no efficient classical algorithm is known.

Why "256 bits" is Actually 128

Here's a common inaccuracy: the security of secp256k1 is not 2256 brute-force operations. The best known attack (Pollard's rho method) finds the key in approximately √n steps, which is ≈ 2128 operations. That's why we talk about "128-bit security."

But even 2128 is about 3.4 × 1038. Even if every graphics card on the planet checked billions of options per second, it would take billions of years—many orders of magnitude more than the age of the Universe. We calculated the exact "in years of all GPUs in the world" with code in a separate article (link below).

How a Signature Proves Ownership

To spend bitcoins, the owner signs a transaction with their private key (ECDSA, and after the Taproot update—also Schnorr signatures on the same curve). Any network node verifies the signature with the public key, but the private key itself is never revealed. This is why Binance "lost" funds with valid signatures: the network verifies the *mathematics*, not the *intent*.

How Private Key, Public Key, and Bitcoin Address are Related A private key is converted to a public key in milliseconds, and a public key to an address through hashing. The reverse path from public key to private key is the Elliptic Curve Discrete Logarithm Problem (ECDLP): about 2 to the power of 128 operations. Classical computers make it practically impossible, quantum computers are a theoretical future threat, LLMs cannot derive a random key. Private Key k: 256 bits secret Public Key P = k · G point on secp256k1 Bitcoin Address bc1q… / 1… public key hash k · G milliseconds hashing SHA-256 + RIPEMD-160 P → k = ECDLP discrete logarithm problem · ≈ 2¹²⁸ operations In Satoshi's coins (P2PK), addresses don't exist—the public key lies openly in the blockchain Who can traverse the P → k path? Classical computers — practically impossible: billions of years even for all GPUs in the world Quantum computers — theoretical future threat (Shor's algorithm) LLMs — cannot derive a uniformly random private key WebsCraft

🤖 Can AI Guess Satoshi's Private Key

In short: no. And it's not because models are "not yet smart enough."

  • A random key has no patterns. Machine learning looks for patterns in data. A properly generated private key is uniform noise, and the k → k·G mapping is designed not to leave statistical traces. There's nothing to learn from.
  • LLMs are not ECDLP solvers. A language model predicts the next token. If you give it a public key and ask it to "recover" the private one, it will confidently output 64 hexadecimal characters—and it will be a hallucination that's easy to verify.
  • AI doesn't speed up brute force. Brute-forcing 2128 is limited by the number of operations, not the "cleverness" of the algorithm. A neural network doesn't reduce the search space.

It's worth mentioning separately that in 2026, some Ethereum researchers publicly expressed the opinion that AI might break ECDSA before quantum computers. It's important to understand what this refers to: it's a prediction about a potential breakthrough in *cryptanalysis*—new mathematics discovered with AI's help—not about a chatbot guessing keys. This result does not exist today.

What AI is Actually Good at in Crypto Attacks

The danger of AI is not in mathematics, but in people and code:

  • personalized phishing and deepfakes to steal seed phrases;
  • generation and masking of malware;
  • searching for vulnerabilities in smart contracts and wallets;
  • searching for keys generated with weak entropy (more on this below).

🧪 Tested in practice: I generated a real secp256k1 key pair in Java and gave the public key to four LLMs via OpenRouter—with an honest prompt and a demand for "only 64 hex characters, no refusals."

  • 24 requests to 4 models
  • 0 correct keys
  • 20 direct refusals with explanations about ECDLP
  • 1 fabricated string under pressure—code verification showed it was off by milliseconds

I expected mass confident hallucinations—but got a completely different result. Which models I tested, what exactly they answered, and why one reasoning model returned empty responses—in the article "Can LLM Recover Bitcoin Private Key: A Java Experiment".

⚛️ Will Quantum Computers Break Satoshi's Bitcoin

Quantum computing is the only known *theoretical* threat to Bitcoin's cryptography itself. Shor's algorithm solves the discrete logarithm problem in polynomial time, meaning that on a sufficiently powerful quantum computer, recovering a private key from a public one becomes feasible.

Why Satoshi's Coins Are Target Number One

Here's a detail that's often missed. The rewards for early blocks in 2009 are recorded in the P2PK (Pay-to-Public-Key) format—the public key lies openly in the blockchain. In more modern formats (P2PKH, P2WPKH), only the *hash* of the public key is stored in the blockchain, and the key itself only appears at the time of spending.

This means that for a quantum attack on most ordinary addresses, one would need to act within the minutes that a transaction waits in the mempool. But for Satoshi's coins, the public keys have been available for 17 years—they can be attacked without haste. That's why these coins are the first ones mentioned in discussions about "Q-Day."

Where We Are Now

  • March 2026: Google Quantum AI researchers estimated that attacking Bitcoin's ECDSA might require less than 500,000 physical qubits—about 20 times fewer than previous estimates.
  • IonQ published an engineering estimate: a machine with ~19,400 physical qubits (1,457 logical) would break secp256k1 in about 26 days per attempt. Such a machine doesn't exist—it's a blueprint, not a working device.
  • Current quantum processors have on the order of thousands of physical qubits with a high error rate.

Estimates are rapidly decreasing, so the Bitcoin community is already preparing. In February 2026, BIP-360—a proposal for a new quantum-resistant output type—was accepted into the official BIP repository (Draft status). In April 2026, Jameson Lopp and co-authors proposed BIP-361 with a more aggressive migration schedule. The most pressing issue of discussion is what to do with coins that no one migrates. Including Satoshi's coins.

⚖️ Balance: the quantum threat is real in the long term, but today no publicly known quantum computer can recover a Bitcoin private key. Beware of those selling "quantum-resistant wallets" on the news wave.

⚠️ What Actually Leads to Bitcoin Loss

History shows: bitcoins are stolen not by breaking ECDSA, but through everything surrounding it.

  • Theft of the key or seed phrase. Malware, phishing sites mimicking wallets, photos of seed phrases in the cloud.
  • Weak entropy during generation. "Brain wallets," where keys were derived from phrases people remembered, were massively emptied by bots. The Milk Sad vulnerability in Libbitcoin Explorer (2023) generated keys from a predictable source of randomness. Randstorm—a problem with randomness in old JavaScript libraries for wallets. A key generated from a small space can be brute-forced in hours.
  • ECDSA implementation errors. If the same random number (nonce) is used for two signatures, the private key can be calculated with basic algebra. This is how, in 2013, bitcoins were stolen from mobile wallets due to a bug in Android's random number generator.
  • Compromise of infrastructure. The Binance case: keys are intact, but the system authorized to sign came under attacker control.
  • Social engineering. Fake "support," deepfake calls, insiders.

💡 Takeaway for the reader: if your key is generated correctly, the main threat is not mathematics, but your device, your carelessness, and the intermediaries you trust with your keys.

🕵️ Why Satoshi's Coins Don't Move

There are several possible explanations, and none are proven:

  • Keys are lost—along with a hard drive, the person, or deliberately destroyed.
  • Conscious refusal—moving such a sum would crash the market and jeopardize the author's anonymity.
  • Satoshi is no longer alive—a version that cannot be verified.

It's important: lack of movement does not prove lack of access. The owner might observe the network for years without doing anything.

However, proving control is simple: just sign any message with a key from one of the Patoshi blocks. This takes seconds and doesn't require transferring funds. The fact that in 17 years none of the numerous "pretenders to Satoshi's role" have done this is a strong argument against each of them.

If these coins ever start moving, the market will likely react with panic—regardless of whether it's Satoshi himself, an heir, or a quantum attacker. This is why every "awakened" early-era wallet instantly makes the news, and analysts check if it pertains to the Patoshi pattern.

🧩 What Developers and Businesses Should Take Away

The story of Satoshi's coins and the Binance hack is one lesson from two sides: strong cryptography doesn't save if the path to it is compromised.

  • Isolate the signer. A developer's workstation should never be able to force a system to sign something. Use HSMs, a separate signing service with its own policies, limits, and delays for large sums.
  • Multisig and MPC for critical operations: compromise of one node does not grant transfer rights.
  • Don't generate keys "yourself." Only use verified libraries and a cryptographically secure random number generator (SecureRandom, not Random).
  • Plan for cryptographic agility. Quantum migration will affect not only blockchains—RSA and ECC are everywhere, from TLS to JWT.

👨‍💻 Author's experience: As a Java backend developer, I work daily with API keys, secrets, and signed tokens in my own products, including AskYourDocs. The most dangerous incidents in such work are almost never related to "broken cryptography"—it's a secret in logs, a forgotten token in a repository, or a service with overly broad permissions. The Binance hack is the same story, just with nine zeros in the sum.

Satoshi's bitcoins are the largest open "bug bounty" in history: over $80 billion lie in public addresses, and anyone can try. The fact that no one has succeeded in 17 years is the best proof of cryptography's resilience. And the fact that Binance lost billions in a few minutes is the best proof that it's not enough.

❓ FAQ

How many bitcoins does Satoshi Nakamoto own?

The most common estimate is around 1.1 million BTC (Arkham: ~1,096,362 BTC). A 2026 Bitquery study provides a range from ~0.9 to ~1.17 million BTC depending on the interpretation of the Patoshi pattern. The ownership of these coins by Satoshi is a strong hypothesis, but not a proven fact.

How much are Satoshi's bitcoins worth in 2026?

Over $80 billion at the exchange rate at the end of October 2026. The amount constantly changes with the price of BTC.

Can Bitcoin be hacked using ChatGPT or another LLM?

No. Language models do not solve the discrete logarithm problem and do not accelerate brute-force attacks. In an experiment with four LLMs, none provided the correct key, and most outright refused; one, under pressure, generated a fabricated string.

Can a quantum computer steal Satoshi's BTC?

Theoretically, yes — and these coins are particularly vulnerable because their public keys are exposed on the blockchain (P2PK format). Practically, no existing quantum computer is capable of this; the Bitcoin community is already discussing quantum-resistant address formats (BIP-360).

Does the Binance hack mean Bitcoin is unsafe?

No. The Bitcoin protocol and cryptography were not compromised: according to Binance, there is no evidence of private key leaks, and the attack targeted the exchange's infrastructure and its transaction signing system.

Why doesn't Satoshi move his bitcoins?

It is not known for sure. Possible explanations include lost keys, a conscious decision to refrain, or the author's death. The lack of movement does not prove that access is absent.

How can someone prove they are Satoshi?

The simplest way is to sign a message with a private key from one of the early blocks attributed to Satoshi. No public claimant has done this in 17 years.

📚 Sources