Today, let's take a look at the Nillion project, which is the 65th launchpool on Binance and also a project on coinlist. It completed community round financing on coinlist in June 2024, raising $14 million. The community cost was $0.4. Now, let's determine what a reasonable price for Nillion is.
Currently, the launchpool mining is ongoing for three days, from March 21 to March 24, and will be listed at 21:00 tonight. Total amount in Launchpool: 35,000,000 NIL (3.5% of the maximum token supply), total token supply: 1,000,000,000 NIL, maximum token supply: unlimited (with an annual inflation rate of 1% according to release type and autonomous governance), initial circulation: 195,150,000 NIL (19.52% of total token supply).
1. Introduction
Nillion's biggest feature mentioned is 'blind computing.' So what is blind computing?
Nillion aims to address these limitations by facilitating data transmission, storage, and computation without decryption, thus ensuring that sensitive information remains private and secure throughout its lifecycle. This approach is known as 'Blind Compute,' which decentralizes trust and extends the use cases of decentralized networks into previously undeveloped blank fields, such as private AI agents, private LLM inference, and other industries requiring secure data. By utilizing advanced privacy technologies (PET) such as multi-party computation (MPC), fully homomorphic encryption (FHE), and Trusted Execution Environments (TEE), Nillion allows data to remain encrypted throughout the entire computational process.
2. Architecture
The Nillion Network architecture consists of two main components: Petnet and nilChain. Nilchain serves as the coordination layer while Petnet is the privacy-enhancing technology network. Petnet is composed of multiple node clusters, each providing different technologies and products, with the long-term goal of allowing anyone to launch their own clusters to meet specific application needs or compete for users and receive rewards. The coordination layer is a blockchain responsible for managing shared resources, such as reward mechanisms, cryptoeconomic stakes, and governance, and supports coordination between clusters.
2.1 Nilchain
nilChain is a blockchain that manages the shared resources of the entire Nillion network. It is built on the Cosmos SDK stack, supporting rewards, cryptoeconomic rights, and governance, as well as achieving inter-cluster coordination. Since its primary purpose is coordination, this chain does not have an open execution environment for running smart contracts.
The main coordination layer of the network, the coordination layer of the Nillion network (referred to as nilChain) is responsible for: (1) managing rewards; (2) payments; (3) cryptoeconomic security; (4) coordination among network clusters.
2.2 Petnet
Petnet allows builders to utilize privacy-enhancing technologies (PET) to store and compute data while keeping it encrypted. nilChain supports payments and incentivizes participation in the Nillion Network.
Petnet consists of a network of nodes, and developers can recruit them into clusters (depending on the PET they use). Developers have the right to choose their own points in the secure computing trade-off space and decide what is important to them.
Petnet nodes support the secure storage and computation of data, which can be leveraged using various SDKs available for interacting with the nodes. Each node supports the use of PET by operating one or more blind modules.
2.3 Blind Modules
Blind modules are the core components of Petnet, and external SDKs are built on top of them. Currently, there are three blind modules: nilVM, nilDB, and nilAI.
2.3.1 nilVM
nilVM is a virtual machine that allows developers to easily create multi-party computation (MPC) programs based on linear secret sharing schemes (LSSS). Programs are written in the Nada language, a domain-specific language (DSL) based on Python, designed similarly to Python libraries, enabling Python developers to get started quickly. Developers can use the Nillion SDK for development, and Nada also provides some auxiliary libraries, such as:
-nada-ai: an interface similar to PyTorch and scikit-learn for AI-related applications.
-nada-numpy: for numerical computation.
-nada-data: for data processing.
2.3.2 nilDB
nilDB is an encrypted, distributed NoSQL database that allows users and developers to unlock the value of sensitive data while protecting privacy. Its core mechanism is to store data in a distributed manner across multiple node clusters, with each node holding only a secret shared portion of the data. Data owners can authorize others to execute SQL-like queries, including computation and privacy-preserving aggregated analysis, making the data stored in the Nillion network more composable and valuable.
The earliest applications developed based on nilDB mainly focus on data marketplaces and insights, leveraging Web3 technology to unlock the massive data generated daily by users through connected devices. This data was originally stored on centralized servers by Web2 private companies, while nilDB not only provides secure storage but also supports data analysis and trading. Whether for AI training data buyers or users seeking aggregated analysis, nilDB makes the use of this data more secure and efficient, while promoting the growth of data storage and computational queries in the Nillion network.
Main Cooperative Projects
-ZAP Labs: founded by former Starkware developers and Kakarot zkEVM co-founders, ZAP is building decentralized data infrastructure centered on zkTLS, storing user data, and providing data analysis services. They initially designed based on Filecoin but ultimately chose Nillion as their storage and computing backend due to its stronger security and programmability.
-Fulcra: founded by the creators of HUMAN (the world's most successful robot defense platform), which was named one of the top 100 most influential companies in the world by TIME in 2023, covering 3 billion devices. Fulcra plans to utilize Nillion to build a smart device data treasury, allowing users to earn through the data market (Data-to-Earn).
-Reclaim Protocol: a zkTLS infrastructure protocol that allows users to prove identity and reputation without trust (e.g., proving bank deposits meet standards without revealing the balance, or proving Uber ride counts without disclosing the itinerary). Nillion, as a storage and computing platform, ensures the private storage and secure processing of this data.
-Healthblocks: helps users collect health data across different devices and platforms, ensuring that ownership and control of the data always remain with the user. Nillion allows insights to be generated without revealing specific information for third-party use.
-Humanity: dedicated to the ownership and monetization of personal data, creating a digital identity and analytics platform that integrates social, health, and entertainment data, ensuring privacy protection while enabling identity verification, cross-platform behavior analysis, macro trend identification, and influence mapping. Nillion, as the underlying infrastructure, provides privacy-preserving data storage and computation.
-Dwinity: dedicated to eliminating information asymmetry and promoting a fair and transparent data economy. They are building a data ownership-driven ecosystem that supports data storage, analysis, exchange, and commercialization. Nillion serves as its core infrastructure to achieve privacy-preserving data storage and processing.
-Verida: developing a personalized AI platform supporting user interaction with AI chatbots trained on personal privacy data. nilDB is used to store personal user data, and AIVM is used for encrypted inference, ensuring user data privacy during AI training and interaction.
2.3.3 Nil AI
nilAI is Nillion's secure AI product that can safely run LLMs within TEE. nilAI is the blind module supporting SecretLLM.
nilAI currently includes AIVM, nada-AI, and nilTEE.
2.3.3.1 AIVM (AI Virtual Machine)
AIVM is a secure AI inference platform based on multi-party computation (MPC), supporting private inference of deep learning models and custom model deployment, maintaining data confidentiality throughout the computation process.
-User inputs and inference results are invisible to the computing nodes, ensuring privacy protection.
-Models can choose to keep user information confidential to prevent model leakage.
AIVM accelerates inference using discrete wavelet transform (DWT), a technology co-researched by Nillion and the Meta AI research team, built on Meta's CrypTen secure machine learning framework.
2.3.3.2 nada-AI
nada-AI is the AI library of nilVM, providing an interface similar to PyTorch, allowing developers to easily run AI models in the Nada DSL language:
-Supports models such as neural networks (NN), convolutional neural networks (CNN), linear regression, etc.
-Provides Colab Notebooks to facilitate quick experimentation and application for developers.
2.3.3.3 nilTEE (Nillion Trusted Execution Environment)
nilTEE is based on Trusted Execution Environment (TEE), meeting developers' needs for efficient inference of large-scale LLMs (large language models):
-Short-term inference is suitable for TEE, avoiding long-term storage of user data and reducing the risk of exposing trust models.
-Nillion is researching 'Split Inference,' which combines AIVM with nilTEE to improve security and performance.
3. Ecosystem
Currently, there are about 35+ projects, including public chains like NEAR, APTOS, ARB, and others such as wallets, payments, e-commerce, and social apps are also starting to build projects on it.
4.Team
Alex Page is the CEO of Nillion. He is an experienced former Goldman Sachs banker and a general partner at Hedera SPV.
Miguel de Vega is a co-founder and Chief Scientific Officer of Nillion, a doctoral advisor with over 30 patents in the fields of machine learning, data optimization, and mathematics. His research background is crucial to Nillion's innovative approaches in privacy-enhancing technology.
Other team members are shown in the figure below, basically all having overseas backgrounds, and all are real-name photos.
5.Financing Scale
Currently, Nillion has completed two rounds of public VC financing.
In December 2022, Nillion completed a $20 million Series A financing round, with a valuation of $400 million. Distributed Global led the investment, with participation from AU21, Big Brain Holdings, Chapter One, GSR, HashKey, OP Crypto, and SALT Fund.
In December 2024, Nillion completed a $25 million financing round, led by Hack VC, with investors including Hack VC, Distributed Global, and HashKey.
6.Token Economics
The total token supply is 1 billion tokens, allocated as follows: 29% for ecosystem and research and development, 21% for early investors, 20% for core contributors, 20% for the community, and 10% for others.
The release rules are shown in the figure below, with core contributors and early investors unlocking some parts after one year, and the other three parts starting to unlock after 6 months, fully unlocking in 48 months.
In conclusion, I am personally not very optimistic about the privacy track, but this project's privacy lies in the data privacy of applications, so compared to previous privacy public chains, at least there is a new concept. Particularly after the incident with a certain executive's daughter unboxing online, public attention to privacy has increased. The traditional WEB2 personal data issue is that data is stored in centralized databases, and user data security relies entirely on the 'conscience' of centralized companies, as even if they sell your data, you have no evidence. Therefore, I believe this project has a good opportunity in the future when individuals increasingly value their data.
The financing scale has also reached $50 million, which is a significant number, so the project quality has been voted by VCs. However, the downside is the VC distribution, and there will still be considerable unlocking pressure in the next 6 months.
So let's calculate what a reasonable token price is. Based on the previous Launchpool, if we calculate with a 1% return on BNB (18 million BNB * $600 / 28 million NIL), it would be around $3.8. With a total of 1 billion tokens, the FDV would be $3.8 billion. Currently, XMR is $4 billion. I believe NIL is definitely better in terms of application and technology than XMR, but XMR is still an established leader. So a reasonable price point is $2, which is a 5x increase from the previous community round of $0.4. Therefore, my suggestion is that the price must be low enough to provide value; the project's fundamentals are still good, and it can be considered one of the more 'valuable' projects recently launched.
