# DePIN Projects Gain Momentum in Web3 Ecosystem
Recent developments in the Web3 ecosystem reveal a surge in decentralized physical infrastructure network (DePIN) projects leveraging blockchain technology to manage decentralized networks. This article examines six major players dominating the DePIN space, analyzing their market capitalization, GPU capacity, primary customer base, and service costs. Data is as of August 13, KST.
In the Web3 DePIN domain, projects such as Render (RNDR), Akash Network (AKT), Aethir (ATH), io.net (IO), Nosana, and Netmind (NMT) are gaining traction, each emphasizing their unique strengths to capture market attention. These projects aim to enhance efficiency in the AI and cloud computing markets.
# Market Capitalization and Fully Diluted Valuation (FDV)
Render boasts the highest market capitalization at $3.5 billion, followed by Akash Network and io.net at $870 million and $500 million, respectively. Nosana trails with the lowest market capitalization of $170 million.
When examining fully diluted valuation (FDV)—the market cap assuming all tokens are in circulation—io.net’s worth escalates to $4.5 billion, followed by Aethir at $3.3 billion and Akash Network at $1.4 billion. A significant gap between market cap and FDV indicates a high volume of tokens yet to be issued. The ratios of FDV to market capitalization are highest in Aethir (11.0), io.net (9.0), Netmind (4.3), Nosana (2.1), Akash Network (1.6), and Render (1.4).
# GPU Infrastructure and Pricing
According to official announcements, Aethir and io.net lead in GPU capacity, with over 40,000 and 22,000 units respectively. Conversely, Akash Network and Nosana have relatively modest infrastructures, with 400 and 500 GPUs each.
Each project targets different key customer segments. io.net and Akash Network focus on AI/ML engineers and enterprises, whereas Render caters to retail clients like 3D artists and engineers.
The DePIN services predominantly utilize NVIDIA’s A100 and H100 GPUs, based on Ampere and Hopper architectures, respectively. The H100 offers superior computational performance for AI and high-performance computing (HPC) tasks compared to the A100, despite higher power consumption, making it highly efficient for cutting-edge AI services.
In terms of cost, io.net and Akash Network offer A100 GPUs at competitive rates of $0.76 and $0.78 per hour, respectively. For H100 GPUs, io.net and Aethir lead with rates of $1.19 per hour, while Akash Network follows at $1.49 per hour.
Both A100 and H100 GPUs are used for high-performance AI/ML tasks, but they differ in technical specifications and applications. The A100 is suitable for existing AI model training and inference tasks and is cost-effective, whereas the H100 is optimized for next-generation AI model training and large-scale parallel processing, ideal for advanced AI model development.
These projects differentiate themselves in terms of GPU capacity, service costs, and targeted customer segments. Aethir and io.net hold advantageous positions with large-scale GPU capacity, benefiting cloud computing and AI/ML enterprises. Akash Network and Netmind, however, position themselves favorably for small to mid-sized customers by offering lower-cost services. As the cloud computing market continues to grow, this competitive landscape is expected to become increasingly fierce.
