For researchers and investors in the Bitcoin ecosystem, one question remains persistent: how many individuals or entities truly own and use Bitcoin? While the entire transaction history of Bitcoin is publicly accessible on its ledger, accurately estimating the number of users within the network is a challenging task.
Traditionally, the number of Bitcoin addresses has been used as a proxy for the number of users or holders. However, this method is flawed. A single user can control multiple addresses, and a single address can hold funds belonging to multiple users, such as in the case of a centralized exchange.
A Better Way to Estimate Bitcoin Holders
To more accurately estimate the number of unique entities holding Bitcoin and understand how this number evolves, advanced statistical methods are required. At Glassnode, a "synthetic" statistical solution has been developed. This approach combines industry-standard heuristics, proprietary clustering algorithms, and advanced data science techniques applied to raw on-chain data.
This methodology allows for the aggregation of addresses into single entities, providing a more accurate upper limit on the number of Bitcoin holders. It is important to note that this method addresses the issue of one entity controlling multiple addresses. However, it does not solve the issue of one address holding funds for multiple users (e.g., an exchange). Therefore, the resulting metric is referred to as "entities," not individual "users."
Key On-Chain Entity Metrics
This research has produced a suite of real-time on-chain metrics available via Glassnode Studio and its API:
- New Entities: The number of unique entities appearing on the Bitcoin network for the first time.
- Entity Net Growth: The net change in the number of unique entities, defined as the difference between new entities and "disappearing entities" (entities whose balance drops to zero).
- Whale Entities: The number of unique entities holding at least 1,000 BTC.
- Active Entities: The number of unique entities that are either a sender or a receiver in a transaction.
- Sending Entities: The number of unique entities acting as senders.
- Receiving Entities: The number of unique entities acting as receivers.
It is important to note that these metric values are variable and may change slightly over time as clustering algorithms refine their analysis. However, mechanisms are in place to ensure highly stable values, with statistical fluctuations kept to a minimum (averaging less than 1%).
Quantifying the Number of Bitcoin Entities
The results of this methodology are revealing. Figure 1 shows the daily number of new addresses compared to the daily number of new entities.
Because the clustering method groups addresses into entities, the number of new entities is only a fraction of the number of new addresses. In 2019, for example, the Bitcoin network saw over 355,000 new addresses per day on average. In contrast, the average number of new entities was just over 100,000, representing only about 28% of the new address count.
The ratio between new entities and new addresses, shown in Figure 2, provides valuable insights. A significant dip in this ratio in early 2018, for instance, indicates a period where activity was driven more by existing users creating new addresses rather than by new users entering the network.
A similar comparison between active addresses and active entities yields a very similar picture, further validating the entity-based approach.
So, How Many Entities Hold Bitcoin?
By analyzing the daily net growth of entities (Figure 4) and creating a cumulative summary of this time series, we can estimate the total number of entities at any given point in time.
Figure 5 compares the total number of addresses with a non-zero balance to the total number of entities. According to this analysis, as of January 2020, the number of entities holding Bitcoin was approximately 23.1 million. This is 18.5% lower than the 28.4 million addresses with a non-zero balance at that time.
Furthermore, the entity count demonstrates a steady, one-way upward trend. Unlike the address count, which saw a sharp decline in early 2018, the entity count has consistently grown, pointing to sustainable network adoption.
Bitcoin's Healthy Growth Trajectory
Comparing the net growth of entities versus addresses (Figure 6) is particularly valuable. The growth in addresses shows much higher volatility and larger positive and negative deviations. Crucially, the daily net growth of entities has almost always been positive. In Bitcoin's entire history, there have only been 21 days where the entity net growth was negative.
This metric clearly indicates that Bitcoin's adoption over the past decade has been both healthy and sustainable. To understand the current distribution of wealth and track large players, view real-time tools that provide these advanced metrics.
The Number of Bitcoin "Whales"
Another crucial entity-based metric is the count of "whales"—entities holding ≥1,000 BTC. This metric provides a more accurate picture than simply counting addresses, as it correctly clusters addresses controlled by a single large holder (note: entities identified as exchanges are excluded from this specific metric).
While the trends for whale addresses and whale entities are similar, the entity-based count provides a clearer and more precise estimation of whale behavior at any specific point in a market cycle.
Distribution Statistics
The distribution of entities reveals a clear power-law relationship. The vast majority of entity clusters (about 96%) still consist of just one address. This finding also sets a strict upper limit on the number of entities; the actual number may be lower due to potential missing links in the clustering process.
Similarly, the distribution of entities by their Bitcoin balance follows a power law. The overwhelming majority of entities (about 22.3 million) hold less than 1 BTC.
As of January 2020, the data showed:
- 75 entities held ≥10,000 BTC.
- 7 entities held ≥100,000 BTC.
All seven of these ultra-large entities were identified as cryptocurrency exchanges: Coinbase (983,800 BTC), Huobi (369,100 BTC), Binance (240,700 BTC), Bitfinex (214,600 BTC), Bitstamp (165,400 BTC), Kraken (132,100 BTC), and Bittrex (118,100 BTC).
Combined, these seven exchanges controlled approximately 2.35 million BTC, equivalent to about 13% of the circulating Bitcoin supply at the time. Figure 10 illustrates the account balance distribution across all entities.
Frequently Asked Questions
What is the difference between a Bitcoin address and a Bitcoin entity?
An address is a single destination on the Bitcoin network. An entity is a cluster of one or more addresses that are controlled by the same network participant, as identified by advanced clustering algorithms. This provides a more accurate estimate of unique holders.
Why is the number of entities lower than the number of addresses?
Because a single user or institution (entity) can control hundreds or thousands of different addresses. Entity-based counting consolidates these addresses into a single unit, preventing overcounting.
How accurate are these entity estimates?
While not perfect, entity-based metrics are significantly more accurate than simple address counts. The clustering algorithms used are industry-standard and provide a robust upper-bound estimate of the number of unique Bitcoin holders.
What defines a Bitcoin "whale"?
In this context, a "whale" is defined as any single entity (not an exchange) that controls a balance of 1,000 BTC or more. This metric helps analysts track the behavior of large, influential holders.
Do these numbers include exchange holdings?
Yes, the total entity count includes exchanges. However, for the specific "whale" metric, exchanges are typically excluded to focus on individual large holders.
How can I track these metrics myself?
On-chain data analytics platforms provide public access to these advanced metrics, allowing anyone to monitor network growth, entity behavior, and whale movements in real-time. You can explore more strategies for analyzing this data on dedicated platforms.