What is Tensorplex Labs? Tensorplex team, financing, product introduction

What is Tensorplex Labs? Tensorplex Labs is building the infrastructure needed for decentralized artificial intelligence, providing a range of solutions including AI-incentive training, open source large language models, liquid staking for $TAO, and native decentralized exchanges within the Bittensor ecosystem.

Below, we will introduce you in detail what Tensorplex Labs is? As well as a detailed introduction to the Tensorplex team, financing and products, let’s take a look at it if you need it!

 introduce

Artificial intelligence is becoming the core driving force for the next wave of technology. Just as we have witnessed the transition of human society from relying on heavy machinery to the digital age of computers and the Internet, the arrival of AI also marks a new transformation. In the process of these technological evolution, control is always concentrated in the hands of governments and large enterprises, who dominate the direction of technology development to achieve their own goals.

Today's AI development is also showing a similar trend, with large technology companies and government agencies generally adopting a closed R&D environment. In contrast, blockchain technology brings the possibility of decentralization, breaking the centralized monopoly and allowing more individuals and organizations to participate in decision-making and innovation.

Tensorplex Labs is one of the pioneering forces in the field of decentralized AI, promoting a development model that is no longer centered on corporate interests but is driven by knowledge sharing and collective progress.

 What is Tensorplex Labs?

 Source: Tensorplex

Tensorplex Labs is a decentralized artificial intelligence infrastructure company dedicated to developing tools and applications for AI development through decentralized networks. The company was founded in 2023 and integrates the concept of decentralized finance (DeFi) and the proof-of-stake mechanism to support innovation in open source artificial intelligence.

Its portfolio includes Dojo, a liquid staking solution for $TAO. At the same time, Dojo is also a platform for collecting high-quality artificial data to accelerate AI training and continue to advance research on training open source large language models (LLM).

 Tensorplex Team

Darwin Liew

Darwin is the co-founder of Tensorplex Labs and was a founding partner of PetRock Capital. He graduated from Nanyang Technological University.

CK Cheung

Cheung is the co-founder of Tensorplex Labs and has worked for Brevan Howard Digital, DeFiance Capital and Goldman Sachs. He graduated from Cornell University and Hong Kong University of Science and Technology.

Calvin Choo

Calvin Choo is the chief software engineer at Tensorplex Labs and worked on the engineering team at a centralized exchange. He graduated from Nanyang Technological University with a bachelor's degree in computer engineering.

 Financing situation

 Source: Tensorplex

In March 2024, Tensorplex Labs completed a $3 million seed round led by Canonical Crypto and Collab+Currency, followed by Digital Currency Group, Quantstamp and Amber Group. YZi Labs will also join its investor list in March 2025.

 Bittensor Blockchain Overview

Bittensor is a decentralized blockchain network that serves as a platform for producing and trading digital goods, including AI reasoning, training and infrastructure services.

It is a peer-to-peer market where participants can contribute computing resources and machine learning models in exchange for rewards. The network uses TAO tokens to incentivize miners (data providers) and validators to improve the intelligence of the system.

The Bittensor blockchain is built on Polkadot's Substrate SDK, runs as a decentralized recording system and uses a custom consensus mechanism - Yuma Consensus. This consensus mechanism is used to reach an agreement between validators and miners in each subnet of the network.

The design of Yuma consensus is metrically independent and can reach an approximate consensus on the basis of non-absolute truth, rather than a rigid binary judgment.

 Bittensor and its subnet

In Bittensor, a subnet is a network segment designed specifically for generating a certain type of AI-related digital goods. For example, one subnet may focus on natural language processing tasks, while the other subnet focuses on computer vision applications. Each subnet operates as an autonomous market and consists of two major players:

Miner: An entity or individual responsible for generating AI-related outputs, such as machine learning models or data processing services.

Verifier: Responsible for evaluating the quality and practicality of miner contributions to ensure that the output meets certain standards.

The interaction between miners and validators in the subnet is constrained by incentive mechanisms, and high-quality contributions will be rewarded.

You can read more about Bittensor in this Gate Learn article.

 Tensorplex Products

 Dojo Subnet

 Source: Dojo

Dojo Network is an open platform where anyone can participate in AI training through human feedback reinforcement learning (RLHF). The network performs machine learning in a decentralized environment through crowdsourcing human-generated datasets.

Participants from different technical backgrounds contribute human intelligence to the operation of machine learning models and AI agents that will be deployed on-chain and in real-world applications.

 Functional Features

Synthetic Task Generation: Dojo uses advanced large language models (LLMs) to generate tasks to collect human feedback. The system learns from these feedbacks and continuously improves the effectiveness of AI training.

Synthetic Ground Truth Validation: This system uses simulated or synthetic data to create benchmarks that evaluate the performance of AI models. This process allows participants’ response quality to be evaluated, ensuring that the collected data meets expected criteria.

Backprop Finance

 Source: Backprop Finance

Backprop Finance is a decentralized exchange (DEX) developed by Tensorplex Labs for trading dynamic TAO (dTAO) tokens. Each dTAO token corresponds to a subnet in the Bittensor network and is unique.

The platform provides a terminal that seamlessly exchanges between TAO and subnet tokens while managing slippage and price shocks.

 Source: Backprop Finance

In addition to the transactional capabilities, Backprop Finance provides access to active subnets, GitHub repository, social links, and exclusive research reports curated by Tensorplex Labs. Users can browse professional-level price charts, real-time positioner analysis, trading history, and portfolio insights. The system also has a ranking mechanism to highlight the best performing subnets and participants.

 Source: Backprop Finance

Liquidity providers can provide liquidity to the fund pool on Backprop and receive dTAO rewards. dTAO holders have governance rights and can vote on which fund pools are awarded dTAO incentives. In return, they can share the fees incurred by the supported fund pool.

Tensorplex Stake

 Source: Tensorplex Stake

Tensorplex Stake is a TAO token liquid staking solution based on the Ethereum network, introducing stTAO - an ERC-20 liquid staking derivative designed to simplify users' participation in the Bittensor ecosystem.

Users can stake TAO tokens through EVM-compatible wallets and obtain stTAO without using a Bittensor-compatible wallet. The wallet integrates Chainlink's Cross-chain Interoperability Protocol (CCIP) and supports cross-chain staking operations on Arbitrum, Base, Ethereum, and Optimism.

 Source: Tensorplex Stake

To encourage user participation, Tensorplex Stake has launched a reward mechanism where users can obtain stTAO Intelligence Chips by:

• Holding stTAO on Ethereum or Base
• Provide liquidity for stTAO/ETH, stTAO/USDC liquidity pools on Uniswap v3 (Ethereum mainnet) or Aerodome (Base chain)
• Cross-chain TAO to Ethereum via Tensorplex Bridge
• Cross-chain TAO from the Bittensor Finney network to Ethereum

stTAO holders may also be eligible for airdrop rewards from Tensorplex collaborative projects.

To use Tensorplex Stake, users can purchase TAO tokens from Gate.io and withdraw them to an EVM-compatible wallet.

TPLX LLM

 Source: Tensorplex

Tensorplex Labs is developing a large language model (LLM) designed for Web3 decentralized applications. Its existing models (tensorplex-labs/Sumo-T9-7B-v0.1) have surpassed the TII Falcon 7B and Meta's Llama-2-7b-hf in performance.

Tensorplex Stream

 Source: Tensorplex

Tensorplex Stream is an AI-powered application that aggregates and summarizes relevant information for cryptocurrency investors. Currently in beta, the platform organizes content from key opinion leaders in the Web3 space and provides the latest dynamic insights from the crypto industry.

Stream is a demonstration application created by the Tensorplex team to demonstrate how to leverage existing Bittensor subnets to support the operation of Web3 consumer platforms.

Tensorplex REP

REP is a community management and customer service chatbot powered by a large language model that can automatically answer questions, manage and review discussion content, and process customer service work orders for Web3 projects.

The impact of Tensorplex on the Bittensor ecosystem

 Ecosystem Growth

As of March 2025, the total lock-in value of Tensorplex's liquid staking platform exceeded US$8 million. Recent investment from YZI Labs triggered a 500% surge in Dojo subnet projects in one day, showing strong market interest in decentralized artificial intelligence.

As Tensorplex continues to optimize its products and establishes strategic partnerships, its influence in the industry is expected to further increase.

Integration of AI and DeFi

Through platforms such as Backprop Finance, Tensorplex is integrating artificial intelligence (AI) with decentralized finance (DeFi); dTAO token trading linked to different Bittensor subnets is expected to promote the birth of new financial products and services, expanding the practicality and accessibility of these two fields.

 Cross-subnet verification and external access

Tensorplex implements a cross-subnet verification mechanism, where validators can evaluate the output of multiple subnets, thereby improving the reliability of the AI ​​model. External entities outside the Bittensor network can also access these subnets to obtain high-quality human-generated data for use in AI application scenarios.

 in conclusion

Tensorplex Labs is building the infrastructure needed to integrate decentralized technologies with artificial intelligence, laying the foundation for a more collaborative and equitable future for AI development. Their innovative products, such as Dojo, make AI training more efficient while attracting a wider range of contributors to create an inclusive community.

Backprop Finance is another platform designed to drive the growth of decentralized finance (DeFi) on Bittensor by creating decentralized exchanges for Bittensor subnet tokens.

As companies like Tensorplex continue to innovate, the next generation of technology seems to be moving in a more selfless direction.