How do big banks and financial funds with 1,000+ Bitcoin store their Bitcoin?

If you've been in the world of Bitcoin for a while, you've probably heard this phrase:

"Not your keys, not your coins."

A short phrase that has become a fundamental principle in the crypto industry for over a decade.

For retail investors, the most widely accepted practice is to hold one's own Private Key via a Hardware Wallet, such as Trezor, Coldcard, BitBox, or other brands, and store the Seed Phrase in a secure location. This concept is called Self-Custody.

But as Bitcoin began to be adopted at an institutional level, new questions arose.

If the Bitcoin holder is not an individual, but a bank, an ETF fund, a public company, or an asset manager overseeing hundreds of billions of baht for clients, how do they store Bitcoin?

The answer might not be what many expect.

Because while the ultimate goal remains to protect the Private Key, the underlying structure is completely different from the Hardware Wallets most investors are familiar with.

This article will take you on a deep dive into this technology.

The real problem isn't storing Bitcoin, but storing the "keys."

Technically, Bitcoin is not stored in a Hardware Wallet; all Bitcoin always resides on the blockchain. What a Hardware Wallet actually stores is the Private Key, a digital key used to prove ownership and authorize asset transfers.

The 12 or 24-word Seed Phrase we are familiar with is merely a human-readable form of the data used to generate the Private Key.

For this reason, the security of Bitcoin does not depend on the coin itself, but on the protection of the Private Key.

For retail investors, this is a relatively straightforward problem: we have one set of keys, and we protect that single set as best as we can.

But at an institutional level, having "only one set of keys" becomes a problem.

Imagine a fund holding 20 billion baht worth of Bitcoin.

The questions that need to be answered are not just:

As soon as assets reach the billion-baht level, security is no longer about devices, but about system design.

Why Hardware Wallets Are Not Enough for Large Organizations

Hardware Wallets are excellent tools for individuals, as they help keep Private Keys separated from the internet, reducing the risk of malware or online attacks.

However, for large organizations, Hardware Wallets have certain limitations.

Firstly, there is a Single Point of Failure.

In general use, one Hardware Wallet often protects one Seed Phrase. If a single individual can access the Seed Phrase or Private Key, it could mean that the organization is entrusting the security of assets worth billions or tens of billions of baht to just one person, which is not aligned with the risk management principles of large organizations.

Secondly, there is the issue of Governance.

Large organizations cannot allow asset movements to occur based solely on the decision of one individual. Transactions worth hundreds of millions of baht may require approval from multiple parties, such as risk management, regulatory affairs, senior executives, or the board of directors.

Hardware Wallets are designed to protect keys (Private Keys) but not to manage organizational approval processes.

The Dawn of Institutional Custody

In the early days, many organizations opted for a system called Multi-Signature Wallet, or Multi-Sig for short.

The concept of Multi-Sig is quite straightforward: instead of one key, the system generates multiple keys. For example, a 3-of-5 wallet means there are a total of 5 keys, but at least 3 are required to authorize a transaction.

If compared to the real world, this is like a safe equipped with 5 padlocks, requiring at least 3 keys to open. The advantage is that no one person can move assets alone.

This concept significantly reduces the risks of both theft and human error.

However, Multi-Sig still has several limitations, such as:

These limitations make Multi-Sig inconvenient for use across various blockchains or within large organizations.

While MPC is currently very popular, it doesn't mean that every institution uses MPC exclusively.

In reality, each Institutional Custody provider may choose to use different architectures, such as Multi-Signature, Hardware Security Module (HSM), MPC, or even a combination of several technologies, to suit their security level, legal requirements, and the blockchains they need to support.

Therefore, the term Institutional Custody doesn't refer to a specific technology, but rather to an approach for designing systems to protect digital assets at an organizational level.

This led to the emergence of a technology that is now very popular in the industry.

That technology is MPC.

What is MPC?

MPC stands for Multi-Party Computation.

Originally, MPC was not invented specifically for crypto, but stemmed from cryptography research dating back to the 1980s. The goal of MPC is to enable multiple parties to jointly compute something without revealing their secret data to each other.

When this concept was applied to digital asset custody, the result was a complete change in how Private Keys are perceived.

Instead of creating a Private Key and then dividing it for storage, MPC creates what are called Key Shares, or "mathematical shares of the key," from the outset. Each share is not a Private Key and cannot be used alone.

Mathematically, the Private Key still exists, but no single party holds the full Private Key. In many modern MPC system designs, the Private Key is never assembled into a single unit during use.

What Makes MPC Different?

The most interesting aspect of MPC is that, in the MPC Custody systems widely used today, many are designed so that the Private Key does not need to be assembled into a single unit during transaction signing.

This is a point that even many crypto investors still misunderstand. Many think that MPC is merely dividing a Seed Phrase into several parts.

But that's not actually the case.

If you divide a Seed Phrase into three pieces, you still have the original key underneath.

Conversely, MPC uses mathematical principles to create Key Shares from the start. Each share cannot be reverse-engineered to find the original key. And when a transaction is signed, the system uses each share collaboratively to create a Digital Signature without needing to reveal or assemble the full Private Key during the process.

Fireblocks, one of the world's largest digital asset custody infrastructure providers, briefly explains this principle as:

"The complete key is never assembled in one place."

In other words, the full key is never gathered in one spot for anyone to steal from the beginning.

What if a Share is Lost?

This is a very important question.

If a Hardware Wallet is lost, we can recover it with a Seed Phrase.

But in a typical MPC system, there is no BIP39 Seed Phrase that users need to write down like with a Hardware Wallet. So what then?

The answer is that most institutional MPC systems are designed with Threshold Policies.

For example, a system might have 5 shares in total, but only 3 shares are needed to create a digital signature.

This means that even if one or two shares become unusable, the system can still operate normally.

In practice, these shares are often distributed across different locations and environments, such as:

Therefore, even if some servers crash, some offices become inaccessible, or some equipment is damaged, the system can still operate according to the defined policy.

Comparison Table: Hardware Wallet vs. Multi-Sig vs. MPC

Do Hardware Wallets still have a role in the world of MPC?

Having read this far, many might start to wonder if Hardware Wallets will become irrelevant if institutions use MPC.

The answer is no.

In reality, these two technologies are not competing but can work together.

Many institutional custody providers opt for a Hybrid architecture, combining MPC with other technologies such as Hardware Security Modules (HSM), Cold Storage systems, or specialized hardware devices to further enhance security.

From another perspective, MPC doesn't replace Hardware Wallets but addresses problems that Hardware Wallets weren't designed to handle, such as multi-party transaction approval, multi-level user privilege management, and mitigating the risk of any single individual controlling all assets.

Therefore, in the world of digital asset custody, both Hardware Wallets and MPC have their own roles, but they are designed to solve different types of problems.

Hardware Wallet vs. MPC: Not a competition

Having read this far, many might start to wonder if MPC is better than Hardware Wallets.

The answer is not necessarily.

Because both are designed to solve different types of problems.

Hardware Wallets are created to enable individuals to securely hold their assets themselves.

MPC, on the other hand, is created to enable organizations to manage assets collaboratively without relying on any single individual.

If you are a retail investor holding Bitcoin worth hundreds of thousands or millions of baht, a Hardware Wallet remains the simplest and most suitable option for Self-Custody.

However, when assets are worth tens of billions of baht and involve multiple parties in their management, security is no longer just about keeping private keys safe. It becomes about designing systems, governance, and carefully distributing control over assets.

For this reason, many institutional custody providers choose MPC or a Hybrid architecture that combines MPC, Multi-Signature, Hardware Security Modules (HSM), and internal control systems to create an asset custody structure suitable for the risks and requirements of each organization.

Ultimately, whether it's a Hardware Wallet, Multi-Sig, or MPC, they all share the same goal: to protect private keys as securely as possible, simply by choosing the approach that suits the asset size and user type.