📊 Bitcoin’s Decentralization: What the Data Really Says
Bitcoin is intended to be a decentralized, trustless peer-to-peer currency. But how decentralized is the network, really, 15 years after its creation?
This question became the cornerstone of my project a Bitcoin Decentralization Dashboard that measures decentralization over time using actual blockchain metrics. Using statistical tools and daily network snapshots, we wanted to try to strip away much of the assumptions and provide a simple but significant piece of data: how decentralized is Bitcoin today?
🧪 Measuring Decentralization
To measure Bitcoin’s decentralization, we rely on three well-established metrics referenced in the literature: the Gini Coefficient, Shannon Entropy, and the Nakamoto Coefficient. These metrics allow us to analyze how evenly distributed key aspects of the network are such as mining power, node distribution, and more.
- Gini: Measures inequality. A value of around 1 indicates most control, concentrated in very few hands.
- Shannon Entropy: Which reflects how much randomness or diversity is in the system higher means greater decentralization.
- Nakamoto Coefficient: The number of entities that control over 50% of the resource. Higher values = healthier decentralization.
The calculations were conducted along different vectors of the network: over the mining pools, the locations of nodes by country and ASN, and the software diversity.
📈 What the Data Shows
Over the 3-week period between September 23 and October 18, 2024, the dataset revealed some interesting things and in some cases of concerning patterns.
1. Mining Power Is Still Highly Centralized
Even though mining hardware has become more accessible over the years, the centralization in mining pools is still stark. The Gini coefficient for mining hovered around 78–82%, indicating a steep imbalance.
- Nakamoto Coefficient: ~16–17 only about 16 mining pools are needed to control 51% of the hash power.
- Shannon Entropy: Averaged around 12.1–12.2%, suggesting limited distribution.
The data confirms a long-standing concern : mining power is concentrated
2. Node Distribution Is Slightly Better, But Not Great
The total number of reachable nodes fluctuated between 18,000 and 19,500, and geographically, the network appears to have spread to multiple countries. However, inequality still exists.
- Gini for Nodes: ~63%–67%
- Shannon Entropy: ~35%–38%
- Nakamoto Coefficient ASN: Often in the 8–10 range, which means a few autonomous systems or regions could exert majority influence.
This suggests that while nodes are better distributed than mining power, the diversity is still limited, especially when looked at through ASN (ISP-level) or protocol versions.
3. Country-Level Distribution Shows Moderate Spread, But Centralization Persists
When analyzing the geographic distribution of Bitcoin nodes, the data suggests a moderate spread across multiple countries but with significant concentration in a few.
- Gini for Country: Typically ranged between 63% and 67%, pointing to noticeable inequality in distribution.
- Entropy: Held steady around 35%–38%, reflecting some geographic diversity but not enough to consider it well-balanced.
- Nakamoto Coefficient (by Country): Often stay between 8 and 10, meaning just a handful of countries host the majority of active nodes.
While this shows better distribution than mining power, the geographic centralization remains a weak point. Concentration in a few countries could pose risks to network neutrality, especially in the face of regional regulations or infrastructure disruptions.
4. Trends Are Surprisingly Stable
One interesting insight from the dashboard is how stable these values are day-to-day. Despite fluctuations in hashrate and difficulty, the decentralization scores remain in tight bands.
This could mean two things:
- The network has reached a sort of centralization equilibrium.
- Or, there’s a lack of new node diversity and mining distribution over time.
Both interpretations are worth deeper investigation.
🧭 Why This Matters
Understanding decentralization isn’t just a theoretical exercise it’s crucial for Bitcoin’s security, resilience, and censorship resistance. A highly centralized network is more vulnerable to:
- 51% attacks
- Regulatory targeting
- Censorship from service providers
By visualizing and measuring decentralization over time, this dashboard can help Bitcoin researchers, node operators, and enthusiasts make informed decisions and maybe even push for improvements.
🛤️ What’s Next?
The current version of the dashboard focuses on Bitcoin’s base layer. But there’s more to explore:
- Lightning Network Decentralization — How do second-layer networks compare?
- Ethereum and other chains — Cross-chain comparison of decentralization trends.
- Forecasting models — Can we predict when decentralization might degrade further?
There’s room to integrate alerts, interactive simulations, and more layers of transparency in future versions.
📁 Project Resources …!
If you’re interested in exploring the data or code behind this project:
- Dataset Collected : https://drive.google.com/file/d/1PWMLmyoEhiHoANDZUFelUCUwqm8SWpR0/view?usp=sharing
- Dashboard Code: Reach out to me via email at chandrakumar.rajakumar@uab.cat for access or collaboration