Internet Computer (ICP) sustainability report

Internet computer is present on the following networks: Base, Ethereum, Internet Computer.

Base operates as a Layer-2 (L2) scaling solution built on the Ethereum blockchain, having been developed by Coinbase using Optimism's OP Stack. Critically, Base L2 transactions do not possess an independent consensus mechanism. Instead, their validation is directly linked to and secured by the underlying Ethereum Layer-1 (L1) network. This is achieved through a specialized component known as a sequencer. The sequencer's role is to aggregate multiple L2 transactions into bundles, which are then regularly published to the Ethereum mainnet as a single L1 transaction.

Consequently, all transactions processed on the Base network are indirectly secured by Ethereum's robust Proof-of-Stake (PoS) consensus mechanism once they are recorded on L1. Ethereum's PoS system, established with "The Merge" in 2022, moves away from energy-intensive mining by requiring validators to stake at least 32 ETH. In this system, a validator is randomly selected every 12 seconds to propose a new block, while other validators on the network are responsible for verifying its integrity. The network employs a sophisticated slot and epoch system, with transaction finality typically occurring after two epochs, which translates to approximately 12.8 minutes, utilizing the Casper-FFG protocol. The Beacon Chain is central to coordinating validators, and the LMD-GHOST fork-choice rule ensures the chain adheres to the path with the most accumulated validator votes. Validators are incentivized with rewards for their participation in proposing and verifying blocks, but face stringent penalties, known as slashing, for any malicious actions or prolonged inactivity. This design choice by Ethereum aims to significantly enhance energy efficiency, security, and scalability, with ongoing and future upgrades, such as Proto-Danksharding, further targeting improvements in transaction processing efficiency, thereby benefiting Base as its foundational security layer. Base specifically leverages Optimistic Rollups as part of the OP Stack, meaning transactions are presumed valid unless challenged within a specified period via fault proofs.

The Ethereum blockchain network, following "The Merge" in 2022, operates on a Proof-of-Stake (PoS) consensus mechanism, a significant departure from its previous Proof of Work system. This transition replaced energy-intensive mining with validator staking, aiming to enhance energy efficiency, security, and scalability. In this model, participants willing to secure the network act as validators by staking a minimum of 32 units of the network's native asset (Ether). The network organizes its operations around a precise slot and epoch system. Every 12 seconds, a validator is randomly selected to propose a new block. Following this proposal, other validators on the network verify the integrity and validity of the block. Finalization of transactions, meaning they become irreversible, occurs after approximately two epochs, which translates to about 12.8 minutes, utilizing the Casper-FFG (Friendly Finality Gadget) protocol. The Beacon Chain plays a central role in coordinating the activities of these validators, while the LMD-GHOST (Latest Message Driven-Greedy Heaviest Observed SubTree) fork-choice rule is employed to ensure all network participants agree on the canonical chain, following the branch with the heaviest accumulated validator votes. Validators are economically incentivized for their honest participation in proposing and verifying blocks, but they also face severe penalties, known as slashing, for malicious actions or prolonged inactivity. This PoS framework is designed not only to reduce the network's environmental footprint but also to lay the groundwork for future upgrades, such as Proto-Danksharding, which are intended to further improve transaction efficiency and overall network throughput. The core components like validator selection, block production, and transaction finality are intrinsically tied to the amount of Ether staked, ensuring that participants have a vested interest in the network's security and stability.

The Internet Computer Protocol (ICP) utilizes a distinctive consensus mechanism that synergizes Threshold Relay with Chain Key Technology, ensuring decentralized, scalable, and secure operations for its underlying network. Unlike traditional Proof-of-Work or Proof-of-Stake systems, Threshold Relay operates by having a designated group of nodes, referred to as "the committee," collaboratively generate a random beacon. This beacon is pivotal for the impartial selection of the subsequent block producer. This protocol is specifically engineered to deliver scalability and high-speed performance while rigorously upholding decentralization, allowing any node to seamlessly integrate into the consensus process. A core facet of Threshold Relay involves a threshold signature scheme, which mandates cooperation among a subset of nodes to forge a valid signature. This design intrinsically guarantees that consensus is attained even when confronted with faulty or malicious nodes, providing robust Byzantine Fault Tolerance (BFT) and safeguarding network integrity.

Complementing this, Chain Key Technology is instrumental in managing the Internet Computer's state, enabling the network to scale effectively across a vast number of nodes while simultaneously providing rapid and secure transaction finality. This innovative technology facilitates the creation and management of numerous independent blockchains, known as subnet blockchains, each operating with its own specialized set of validators. Chain Key Technology empowers the Internet Computer to efficiently support billions of smart contracts without sacrificing speed, primarily by streamlining quick communication across these subnets and fostering comprehensive cross-chain interoperability. Furthermore, the network integrates Canister Smart Contracts, a decentralized model where the computational tasks of these contracts, which encapsulate application logic, are distributed across various network nodes. These canisters are designed to operate autonomously and scale dynamically with the network’s expansion. The combined consensus mechanisms guarantee transaction finality, meaning that once a block is validated and added, it becomes irreversible, providing the high level of security essential for demanding, high-stakes applications.

Internet computer is present on the following networks: Base, Ethereum, Internet Computer.

The Base blockchain, as an Ethereum Layer-2 solution utilizing Optimistic Rollups from the OP Stack, implements incentive mechanisms primarily focused on optimizing transaction costs and ensuring secure asset transfers, leveraging the economic security of its underlying Ethereum L1. A core incentive to use Base is its efficiency in reducing transaction expenses. This is achieved by a sequencer that bundles numerous L2 transactions together, submitting them as a single, consolidated L1 transaction to Ethereum. This process significantly lowers the average transaction cost for individual L2 operations, as the collective L2 transactions share the cost of the single L1 transaction fee, thereby making Base a more economically attractive option compared to direct L1 usage.

For the secure movement of crypto-assets between Base and Ethereum, a specialized smart contract on the Ethereum network is employed. Since Base, as an L2, does not manage its own consensus for fund withdrawals, an additional mechanism is in place to guarantee that only legitimate funds can be moved off the L2. When a user initiates a withdrawal request on Ethereum's L1, a predetermined challenge period begins. During this window, any other network participant has the opportunity to submit a "fault proof" if they detect a fraudulent withdrawal attempt, triggering a dispute resolution process. This entire system is strategically designed with economic incentives to encourage honest behavior and deter malicious activities, although specific details of these economic incentives for fault proof submission are not explicitly outlined beyond the general principle.

Furthermore, Base inherits and benefits from the robust incentive structure of Ethereum’s Proof-of-Stake (PoS) system, which indirectly secures Base transactions. Ethereum validators, by staking a minimum of 32 ETH, are rewarded for proposing and attesting to valid blocks, as well as for participating in sync committees. These rewards are distributed through newly issued ETH and a portion of transaction fees. Under the EIP-1559 fee model, transaction fees comprise a base fee, which is algorithmically burned to manage supply, and an optional priority fee (or 'tip') paid directly to validators. To maintain network integrity, validators face economic penalties, known as slashing, if they engage in malicious conduct or fail to perform their duties. This comprehensive incentive framework ensures strong security alignment for Base by reinforcing reliable validator behavior on its underlying L1.

The Ethereum network's Proof-of-Stake (PoS) system is underpinned by a robust framework of incentive mechanisms and applicable fees, meticulously designed to secure transactions and encourage active, honest participation from validators. Validators, who are essential for the network's operation, commit at least 32 units of the network's native asset (Ether) to secure their role. Their primary incentives include rewards for successfully proposing new blocks, attesting to the validity of other blocks, and participating in sync committees, all of which contribute to the network's integrity and consensus. These rewards are distributed in newly issued Ether, alongside a portion of the transaction fees generated on the network. A key feature of Ethereum's fee structure is the implementation of EIP-1559, which divides transaction fees into two main components. The first is a base fee, which is automatically burned, effectively reducing the overall supply of Ether over time and potentially introducing a deflationary aspect, especially during periods of high network activity. The second is an optional priority fee, also known as a "tip," which users can choose to pay directly to validators to incentivize faster inclusion of their transactions into a block. This dual-fee structure aims to make transaction costs more predictable for users. To enforce honest behavior and prevent malicious activities, the network employs a strict system of economic penalties, including slashing. Validators who engage in dishonest acts or demonstrate extended periods of inactivity risk losing a portion of their staked Ether, providing a powerful deterrent against misconduct and ensuring the long-term security and reliability of the network. This comprehensive system aligns the economic interests of validators with the overall health and security of the Ethereum blockchain.

The Internet Computer Protocol (ICP) employs a comprehensive suite of incentive mechanisms and fee structures designed to engage and reward various network participants, including validators, node operators, and canister developers, thereby maintaining the network’s security, integrity, and operational efficiency.

Validators are compensated for their crucial role in upholding network integrity and security. They stake ICP tokens, receiving rewards for validating blocks, participating in consensus, and ensuring the decentralized network's performance. Similarly, node operators, who are responsible for the physical infrastructure and computational resources supporting the Threshold Relay consensus, are also rewarded for their contributions. Canister developers are incentivized not only through the creation of decentralized applications (dApps) but also by potentially sharing in transaction fees generated from the usage of their dApps and the deployment of smart contracts. Users of these dApps or canisters contribute through usage fees, typically paid in ICP tokens, with developers receiving a share based on application activity.

Governance on the Internet Computer is facilitated by the Network Nervous System (NNS), where ICP token holders can stake their tokens to participate in critical protocol decisions, such as network upgrades, incentive adjustments, and fund allocation. This active participation is rewarded, granting token holders influence over the network's future direction. Staking ICP tokens in the NNS also provides token holders with passive income, further securing the network and aligning their economic interests with its long-term stability.

Regarding applicable fees, every interaction with a canister, or smart contract, on the Internet Computer incurs a transaction fee, generally paid in ICP tokens. These fees are scaled based on the complexity and computational resources consumed by the canister call or network operation. Additionally, developers and users deploying applications are subject to storage fees for canister data, ensuring efficient resource utilization and preventing wasteful storage. Participation in NNS governance, such as submitting proposals or voting, may also involve minor fees to prevent spam and ensure equitable governance. Lastly, node operators, while incurring costs for hardware and operations, have these expenses partially offset by the rewards they receive for providing essential network resources.

Internet computer is present on the following networks: Base, Ethereum, Internet Computer.

The energy consumption calculation for the Base blockchain network is meticulously performed using a "bottom-up" approach, where individual nodes are identified as the primary contributors to the network's overall energy footprint. This methodology is based on empirical data collected from a variety of sources, including publicly available information sites, dedicated open-source crawlers, and proprietary in-house crawling tools. The fundamental aspect of estimating hardware usage within the network involves determining the minimum requirements necessary to operate the client software. The energy consumption profiles of the specific hardware devices identified are obtained from measurements conducted in certified test laboratories, ensuring a high degree of accuracy in these foundational figures.

In the process of calculating network energy consumption, the Functionally Fungible Group Digital Token Identifier (FFG DTI) is utilized when available, serving to identify and encompass all relevant implementations of a crypto-asset within the scope of analysis. These mappings are regularly updated, drawing on data provided by the Digital Token Identifier Foundation. However, the source documents do not provide specific URLs for the public information sites, open-source crawlers, or the Digital Token Identifier Foundation, preventing direct external linking within this summary.

The methodology also incorporates assumptions regarding the hardware deployed and the number of participants operating within the network. These assumptions are rigorously verified with "best effort" against empirical data to ensure their realism and accuracy. A key underlying principle is the assumption that network participants generally act in a "largely economically rational" manner. Furthermore, to adhere to a precautionary principle, conservative estimates are applied in situations of uncertainty, leading to higher projected impacts to mitigate underestimation risks. For a specific token on Base, a fraction of the network’s total energy consumption is attributed, based on the token's activity within the network.

The methodology for calculating the Ethereum network's energy consumption primarily employs a "bottom-up" approach, which focuses on the energy demands of individual nodes that are central to the network's operation. These nodes are considered the fundamental factor driving the network's overall energy use. The assumptions underpinning these calculations are derived from empirical data gathered through a variety of sources, including public information sites, open-source crawlers, and proprietary in-house crawlers developed for this purpose. A critical step in this methodology involves determining the hardware used within the network, primarily by assessing the computational and other requirements necessary to run the client software. The energy consumption characteristics of these identified hardware devices are then rigorously measured in certified test laboratories to ensure accuracy. When quantifying the energy consumption for the network, the Functionally Fungible Group Digital Token Identifier (FFG DTI) is utilized, when available, to identify all implementations of the asset in scope, with mappings regularly updated based on data from the Digital Token Identifier Foundation. The information regarding the specific hardware deployed and the total number of participants in the network relies on assumptions that are diligently verified using empirical data whenever possible. Generally, participants are presumed to act in an economically rational manner. Furthermore, adhering to a precautionary principle, if there is any doubt in estimations, conservative assumptions are made, meaning higher estimates are used for potential adverse impacts to ensure a comprehensive and cautious assessment of energy consumption.

The Internet Computer network's energy consumption is determined through a meticulous 'bottom-up' approach, which considers nodes as the primary contributors to the network's overall energy footprint. This methodology relies on empirical findings derived from an array of public information sites, alongside both open-source and proprietary in-house crawlers, to gather essential data. A crucial step involves estimating the hardware utilized within the network, with the main determinants being the specific requirements for running the client software. The energy consumption profiles of these identified hardware devices are precisely measured in certified test laboratories to ensure accuracy.

To standardize and comprehensively identify all relevant implementations of a blockchain asset within scope, the Functionally Fungible Group Digital Token Identifier (FFG DTI) is employed whenever available. The mappings for this identifier are regularly updated, drawing on data provided by the Digital Token Identifier Foundation. The underlying assumptions regarding the hardware deployed across the network and the total number of participants are rigorously verified using empirical data, with a general presumption that participants act largely as economically rational actors. In scenarios where uncertainty arises, a precautionary principle is applied, leading to conservative (higher) estimates for potential adverse environmental impacts. When calculating the energy consumption attributed to a specific asset or function on the Internet Computer, the overall energy consumption of the entire network is first established. Subsequently, a proportional fraction of this network energy is attributed based on the activity level of the particular crypto-asset or application within the Internet Computer blockchain network.

Internet computer is present on the following networks: Ethereum, Internet Computer.

To ascertain the proportion of renewable energy utilized by the Ethereum network, a specific set of methodologies is applied. The initial step involves pinpointing the geographical locations of the network's nodes. This crucial geo-information is gathered through various means, including publicly available information sites, as well as both open-source and internally developed crawlers designed to scan the network. In instances where comprehensive geographical data for nodes is not directly accessible, the analysis resorts to leveraging "reference networks." These are comparable networks chosen for their similar incentivization structures and consensus mechanisms, providing a proxy for node distribution. Once the geo-information is established, it is then integrated and cross-referenced with public data obtained from "Our World in Data." This comprehensive dataset offers insights into the energy mixes and renewable energy penetration across different regions globally. The final calculation of energy intensity is defined as the marginal energy cost incurred for processing one additional transaction on the network. This approach allows for an estimation of the energy footprint associated with scaling the network's transactional volume. For detailed information and the underlying data sources on the share of electricity generated by renewables, relevant information can be found through sources such as Ember (2025) and the Energy Institute - Statistical Review of World Energy (2024), with further processing by Our World in Data, accessible via Share of electricity generated by renewables – Ember and Energy Institute.

To accurately ascertain the proportion of renewable energy utilized by the Internet Computer network, a comprehensive methodology is employed, beginning with the geographical identification of network nodes. This process involves leveraging a diverse range of data sources, including publicly accessible information sites, alongside both open-source and internally developed crawlers, to pinpoint the physical locations where nodes operate. In instances where specific geographic data for nodes is unavailable, the methodology incorporates a fallback mechanism: reference networks are selected and analyzed. These reference networks are chosen based on their comparability in terms of incentivization structures and consensus mechanisms, providing a proxy for energy source assessment.

Once the geographic information is compiled, it is meticulously integrated with extensive public data sourced from Our World in Data, a prominent resource for global statistics and research. This integration allows for a robust estimation of the renewable energy mix at the identified node locations or inferred locations. Furthermore, the energy intensity of the network is quantified, which is defined as the marginal energy cost incurred for processing one additional transaction. This metric offers insight into the efficiency and energy footprint associated with individual network operations. The public information from Our World in Data, which plays a critical role in this calculation, draws upon data from sources such as Ember (2025) and the Energy Institute's Statistical Review of World Energy (2024). This dataset, specifically the “Share of electricity generated by renewables – Ember and Energy Institute,” is accessible via Our World in Data.

Internet computer is present on the following networks: Ethereum, Internet Computer.

The methodology for determining the Greenhouse Gas (GHG) emissions of the Ethereum network closely mirrors the approach used for energy consumption, focusing on identifying emission sources and their quantification. The initial and fundamental step involves precisely identifying the geographical locations of the network's operational nodes. This data collection is facilitated through a combination of publicly available information, as well as specialized open-source and proprietary crawlers designed to actively discover and map node distributions across the globe. Should there be an absence of specific geographic information for the nodes, the analysis intelligently defaults to utilizing "reference networks." These are carefully selected networks that exhibit comparable characteristics in terms of their incentivization structures and consensus mechanisms, providing a basis for estimating the geographic spread when direct data is unavailable. This collected geo-information is then meticulously integrated with publicly accessible data from "Our World in Data." This integration allows for the application of regional carbon intensity factors to the estimated energy consumption, thereby enabling the calculation of associated GHG emissions. The overall GHG intensity is quantified as the marginal emission generated per additional transaction processed on the network, offering a metric for the environmental impact of increased network activity. For detailed information and original data regarding the carbon intensity of electricity generation, sources include Ember (2025) and the Energy Institute - Statistical Review of World Energy (2024), processed by Our World in Data, available at Carbon intensity of electricity generation – Ember and Energy Institute. This resource is licensed under CC BY 4.0.

The methodology for determining the Greenhouse Gas (GHG) emissions attributable to the Internet Computer network is closely aligned with the approach used for energy consumption, focusing on identifying node locations and integrating broad energy data. The initial step involves locating the geographic distribution of the network's nodes, a task accomplished through the diligent use of public information sites, complemented by both open-source and proprietary in-house crawlers. This geographical mapping is crucial for assessing the localized energy mix that powers the network’s operations. Should direct geographical data for the nodes be unattainable, the methodology prescribes the use of reference networks. These networks are carefully selected for their structural similarities, particularly in their incentivization frameworks and consensus mechanisms, allowing for a comparative analysis of their emissions profiles.

After gathering and inferring node locations, this geo-information is subsequently merged with comprehensive public data from Our World in Data. This integration provides the necessary context to estimate the carbon intensity of the electricity consumed by the network's infrastructure at these locations. A key performance indicator calculated from this data is the GHG intensity, which quantifies the marginal emissions associated with processing a single additional transaction on the network. This metric provides a granular view of the environmental impact per operation. The data from Our World in Data, essential for these calculations, is drawn from esteemed sources such as Ember (2025) and the Energy Institute's Statistical Review of World Energy (2024). The specific dataset utilized is “Carbon intensity of electricity generation – Ember and Energy Institute,” which is made available under a CC BY 4.0 license and can be accessed through Our World in Data.