Aethir’s decentralized Graphics Processing Unit (GPU) network initially started with a focus on cloud gaming, but according to co-founder Mark Rydon, the company has shifted its emphasis towards artificial intelligence. This field is experiencing a massive surge in demand for computational power, which is boosting the worth of GPUs and creating a clash with U.S.-China geopolitical dynamics.
Established in 2022, Aethir has garnered significant attention within the Decentralized Physical Infrastructure (DePIN) sector due to its practical applications in the real world.
DePIN protocols are networks where people collaborate to offer a service in exchange for a token.
Over the past year, the space sector has been exceptionally vibrant, as DePIN coins have garnered a staggering market capitalization of $33 billion according to CoinGecko. Remarkably, DePIN tokens are emerging in a wide range of fields, from flight tracking data to mapping, numbering hundreds for each category.
One of the notable initiatives within the field of decentralized computation, Aethir demonstrates the capability of Artificial Intelligence to run on distributed computing networks, rather than relying solely on massive, centralized data centers.
Aethir’s journey
As a researcher specializing in this field, I must note that initially, Graphics Processing Units (GPUs) were primarily geared towards gaming enthusiasts. Leading companies such as Nvidia and AMD were locked in a fierce competition to create more potent GPU chips. Their ultimate goal was to provide gamers with an immersive experience of higher fidelity and resolution within their virtual environments.
Approximately in the year 2006, scientists working at Nvidia found an interesting overlap: the powerful ability of Graphics Processing Units (GPUs) to handle multiple tasks simultaneously, which enhances graphics performance, also proved beneficial for processing vast amounts of data, thereby speeding up traditional computing tasks.
Join us for a remarkable event hosted by Consensus Hong Kong! Don’t miss out on the premier Web3 and Digital Assets gathering happening Feb. 18-20. Register now using the code CoinDesk15 to secure a 15% discount!
However, it was the year 2006, and technologies like Artificial Intelligence were not yet widely recognized or popular. Despite this, Nvidia continued to be traded at a relatively low price point.
Today, Nvidia is vying with Apple for the position of the globe’s most valued corporation. Over the past few years, Graphics Processing Units (GPUs) have emerged as the spine of the artificial intelligence (AI) revolution, powering a wide range of applications, including innovative AI models and sophisticated machine learning algorithms.
Just as Nvidia initially focused on utilizing GPUs for enhancing cloud gaming services and later found greater profitability in Artificial Intelligence (AI), Aethir started by leveraging idle GPUs from data centers to strengthen cloud gaming platforms. However, as co-founder Mark Rydon shared in an interview with CoinDesk, they eventually recognized that the significant potential for revenue lay in AI.
According to Rydon, Aethir was initially conceived within the realm of gaming. However, it wasn’t long before we understood that the robust GPU cloud we were developing had significant implications for the AI industry. Customers in the AI runtime sector are willing to pay substantially more than gamers, and the need for compute power in AI is simply staggering.
It’s truly remarkable. According to TrendForce, a research company, the AI server market is projected to reach an astounding $205 billion. What makes this figure even more impressive is that they predict the total value of all servers to exceed $300 billion.
Aethir functions as a platform that facilitates GPU computing, making idle GPUs available to companies seeking flexible, on-demand capacity, bypassing the burden of ownership and maintenance associated with having their own hardware.
This is a significant advantage for researchers who aren’t affiliated with elite institutions. They can tap into powerful computing resources without having to spend vast amounts on purchasing their own infrastructure.
In a recent interview at Devcon in Bangkok, Rydon stated that the compute we’re working with is incredibly costly and inefficient management of this resource could lead to significant financial losses. However, our decentralized model enables us to expand into areas where there’s high demand for computational power but limited resources to afford it.
Know your customer, know your compute
Due to Nvidia’s dominance in artificial intelligence technology, Chinese researchers are keen to acquire these highly advanced chips, as they surpass by several generations the power currently accessible domestically.
However, geopolitical issues have interfered. To maintain a competitive edge against Beijing, the U.S. Department of Commerce has imposed a ban on selling Nvidia’s top-tier chips like the H100 to China.
By August 2024, I found myself reading in The Wall Street Journal about Chinese companies circumventing U.S. export controls. They were achieving this by leasing decentralized GPU computing power overseas, a strategy reminiscent of services provided by firms like Aethir.
As a researcher, I must clarify that our high-performance computing resources are not accessible to regions such as China due to specific restrictions. This is particularly true for H100 systems, which are not designed for use in China. To ensure compliance with these regulations, we employ geofencing technology and do not permit the onboarding of such compute resources in China.
Regarding the ongoing shift toward decentralization in the Web3 landscape, Rydon clarified that Aethir doesn’t offer complete freedom without approval or restrictions.
“This is necessary for $100 million deals to happen,” Rydon said.
A system should exist on the Web2 level to monitor service contracts outlining performance and dependability expectations for corporate clients, along with Know Your Customer (KYC) procedures, which confirm the identities of involved parties to adhere to regulatory requirements.
He stated, “It seems that nobody has figured out yet how to construct hardware networks that are completely permission-free. This poses a significant threat to a company’s profitability because it allows others to join and leave the network at will without any consequences.
Read More
- SUI PREDICTION. SUI cryptocurrency
- „People who loved Dishonored and Prey are going to feel very at home.” Arkane veteran sparks appetite for new, untitled RPG
- LDO PREDICTION. LDO cryptocurrency
- Destiny 2: A Closer Look at the Proposed In-Game Mailbox System
- Clash Royale Deck Discussion: Strategies and Sentiments from the Community
- Jennifer Love Hewitt Made a Christmas Movie to Help Process Her Grief
- ICP PREDICTION. ICP cryptocurrency
- Naughty Dog’s Intergalactic Was Inspired By Akira And Cowboy Bebop
- Critics Share Concerns Over Suicide Squad’s DLC Choices: Joker, Lawless, and Mrs. Freeze
- EUR IDR PREDICTION
2025-01-08 18:15