AI and cloud computing are the foundational narratives of technology in this era. In early September, due to Oracle's AI/cloud contract performance exceeding expectations, Larry Ellison's net worth even briefly surpassed that of Elon Musk, making him the world's richest person. The strong demand for high-performance computing and cloud services in the AI industry has also intensified competition among upstream GPU computing power providers. In the future, GPU computing power providers that can offer cheaper, higher-quality computing resources and serve more small and medium-sized enterprises are bound to succeed in the market and gain the favor of a wide range of investors.
In this direction, the distributed GPU cloud computing provider Aethir is at the forefront of the industry. Aethir is committed to aggregating enterprise-grade GPUs into a globally shared computing network, providing high-quality and lower-cost computing resources for AI and cloud gaming companies in need. Currently, Aethir has deployed over 435,000 high-performance GPU containers globally, covering 93 countries and regions, serving more than 150 enterprises and clients.
In the secondary market, Aethir is gradually being value-discovered. Throughout September, the price of ATH continued to rise, with a gain of over 63% for the month, hitting a new high since February 2025. ATH also suffered a heavy blow during "10.11," falling to $0.22, but the price subsequently recovered and temporarily stabilized around $0.045. The sharp decline in the crypto space did not change Aethir's fundamentals; it remains undervalued. As traditional markets gradually understand the advantages of decentralized cloud computing power, and global regulators no longer adopt a one-size-fits-all attitude towards the crypto industry, clarifying the essential differences between utility tokens and securities, Aethir may grow into the next "Oracle."
Value Support Behind Aethir
In Web3, the success of a project requires not only a good narrative but also Product-Market Fit (PMF), which, from a commercial and long-term perspective, even determines the life or death of a project. For a project, short-term market FOMO caused by the narrative will eventually dissipate. Without PMF support and a stable, sustainable business model, projects often end up on the path where token issuance is the finish line.
Aethir is one of the few projects in the crypto space that combines an attractive narrative with PMF. Aethir does not rely on "selling tokens" to make money; its main revenue comes from enterprise clients like AI and cloud gaming companies that have stable demand for high-performance computing resources. Therefore, Aethir's revenue capability does not depend on the bull or bear cycles of the crypto market, which also becomes an important value support behind its token price.
Although the narrative focus in Web3 this year has shifted to DAT and RWA, Aethir still delivered strong revenue performance. Aethir's Q3 2025 revenue exceeded $39.8 million, a 22% increase quarter-over-quarter, with a monthly average revenue of $13.29 million, setting a new company record. The revenue/market cap ratio is 14.46% (industry average: 8–23%).
From an Annual Recurring Revenue (ARR) perspective, according to Aethir's GPU Dashboard data, its ARR has reached $166 million to date, a 13-fold increase compared to the same period last year. This proves that Aethir's distributed GPU cloud computing subscriptions have qualitatively leaped this year, and its customer base is continuously expanding.
According to Aethir's disclosure, in the first half of this year, the number of its partners and clients exceeded 150, including several Web2 innovative AI companies such as Kluster.ai, Attentions.ai, and Mondrian AI. Currently, Aethir's computing power infrastructure covers 93 countries and regions, with over 435,000 high-performance GPU containers deployed, covering mainstream models like NVIDIA H100, H200, B200, and A100, capable of meeting the needs of enterprise clients at different development stages.
Meanwhile, according to Aethir's GPU Dashboard data, Aethir's distributed GPU cloud computing power reaches 32,914,556 TFlops, equivalent to owning over 490,000 NVIDIA H100s, worth hundreds of billions of dollars. In terms of available GPU scale alone, Aethir can already compete with the few global giants capable of large-scale LLM training and high-concurrency inference, ranking among the "global top enterprise-grade" computing power.
DePIN Competitive Advantage: SLA Service Level Agreement Stability Reaches 99.31%
"Aethir is currently the only DePIN project in Web3 that can provide stable services for enterprise-level clients," said Mark, Co-founder of Aethir, when discussing Aethir's competitive advantages in the DePIN track.
Mark's confidence stems from the fact that, in addition to genuinely possessing enterprise-grade computing resources, Aethir also has an SLA. In Q3 2025, Aethir's SLA stability reached 99.31%. An SLA is a written commitment between the service provider and the client regarding service availability, performance, response, support, compensation, and other elements. It clearly states "what level of service the client expects" and stipulates how compensation or handling will occur if the service fails to meet the agreement.
An SLA is not just marketing rhetoric. For a company providing enterprise-grade services, having an SLA means the company has formalized service quality, response, and compensation rules into a formal commitment to the client. This is both proof of trust and professionalism and a legal/financial and operational responsibility. Enterprises also prioritize selecting service providers with SLAs.
In the traditional B2B field, SLAs are almost standard. However, in the emerging industry of DePIN and decentralized cloud computing power, unstable service quality and poor performance are widely criticized by the market. Aethir is the first decentralized cloud computing power service provider to have an SLA. This not only means that Aethir has overcome the shortcomings of DePIN, with its business model becoming mature and systematic, but it also provides Aethir with a significant competitive advantage.
"Aethir does things very differently from other DePIN projects. Io.net primarily uses distributed GPU resources to solve some simple tasks in AI or machine learning; Render provides services for 3D rendering targeting consumer-level computing connections; Grass uses distributed network resources to earn points; while Aethir focuses on providing enterprise-level services at the software layer for areas like gaming and AI," explained Mark, Co-founder of Aethir, regarding the differences between Aethir and other DePIN projects.
A complete business model has also made sustainable revenue from enterprises Aethir's core advantage. Aethir's ARR has reached $166 million, while according to DePIN Pulse data, Render's ARR is only $30.44 million, and io.net's ARR is only $11.31 million. Meanwhile, from Q3 2024 to the present, Aethir's protocol revenue has grown steadily, unaffected by crypto market cycles.
To date, Aethir has generated $115.9 million in revenue, with its revenue over the past 30 days reaching $12.8 million, surpassing Ethereum and Lido.
Aethir's Pricing Advantage and the Trend of Computing Power Financialization
Aethir's real business competitors are Web2's centralized computing and cloud service enterprises, and its biggest advantage is competitive pricing. "For example, right now, the prices on our platform are about 1/10th of AWS's prices." As shown in the chart below, Aethir's hourly GPU quotes are significantly lower than those of Google, AWS, and Oracle, yet it同样具备大规模LLM训练和云游戏渲染的企业级计算资源条件。
This is mainly due to the advantages of Aethir's distributed network. The decentralized operational model allows Aethir to avoid purchasing GPU hardware equipment, bearing data center construction and operational costs, and corresponding corporate management expenses, reducing its costs by 80% compared to traditional Web2 computing power providers. In traditional markets, these costs are ultimately passed on to enterprises purchasing/renting computing power. In Aethir's network, enterprises connect directly with computing power providers, not only reducing intermediary costs but also promoting fair competition, allowing small and medium-sized or startup enterprises to obtain the computing resources they need at lower prices.
Aethir's main service targets are Web2 AI and gaming enterprises, but "the wine is good, yet the lane is deep." Although decentralized computing power has pricing advantages, some enterprises might fear risks due to a lack of understanding of how decentralized cloud computing power operates. Furthermore, from a compliance perspective, Aethir's model of using cryptocurrency for service payments also touches upon corporate compliance boundaries. For实体企业, because the definition and regulatory attitude towards cryptocurrency vary from country to country, using cryptocurrency to settle service payments may involve issues such as anti-money laundering, payment licenses, and taxation.
This is essentially the fundamental difference between "tokenized pricing" and "computing power financialization." In traditional computing power markets, service payments are generally settled in fiat currency between parties. Even if tokens are used internally for pricing, those tokens cannot circulate on the secondary market. However, in the Aethir network, the platform token ATH is used to settle service payments, and ATH is a volatile cryptocurrency that can be freely traded on the secondary market. This was previously the largest regulatory gray area in the decentralized cloud computing power field and has created resistance for Aethir's business expansion.
Fortunately, with the relaxation of US regulations towards the crypto industry this year, and the SEC gradually clarifying the boundaries between securities and crypto utility tokens, AI enterprises' acceptance of decentralized computing power is gradually increasing. Although Web2 giants still have disagreements on whether computing power needs to be provided centrally, they have also explored using cryptocurrency for service payments. In 2022, Google piloted allowing some customers to pay for cloud services with cryptocurrency via Coinbase.
Simultaneously, against the backdrop of the DAT-filled era, traditional finance's acceptance of decentralized computing power tokens has also increased. Aethir is also at the forefront of the industry in terms of computing power financialization. On September 29, Aethir announced a partnership with Predictive Oncology (NASDAQ: POAI) to launch a $344 million ATH Digital Asset Treasury (DAT). This treasury is the first strategic reserve primarily backed by GPU computing power.
Aethir is rapidly expanding its GPU infrastructure from Q4 2025 to 2026, recruiting new cloud hosts globally to meet the explosive growth in enterprise AI demand. By Q1 2026, Aethir plans to more than double its global computing power scale, further consolidating its position as a leading global GPU cloud service provider driving the AI economy.
The actual revenue generated will be used to repurchase ATH tokens. Aethir initiated a $45 million DAT buyback plan on October 11, with a duration of 45 days.
Therefore, Mark called Aethir DAT a work of art, representing the true integration of capital markets and token economics, perfectly synchronized.
It is believed that in the near future, as the role of decentralized cloud computing power in promoting the development of AI enterprises becomes more prominent, global regulators will actively formulate relevant policies to promote the healthy development of the industry. Then, as the leader in decentralized computing power, Aethir's value will be re-evaluated by capital.
ATH's Positive Flywheel
The success of a Web3 project involves not only narrative and PMF but also sound tokenomics. Aethir's tokenomics create a positive flywheel for ATH, becoming the long-term value support for the ATH token price.
Firstly, as the credential for using Aethir's computing power, ATH is deeply integrated into the operation of the Aethir network. On one hand, enterprise clients need to use ATH tokens to settle payments for GPU computing resources, giving the token utility and creating a fee sink system similar to public blockchains. However, unlike them, AI model training is a long-term process, and the demand for computing power and its stability are also long-term, meaning that each client represents a long-term commitment to purchasing ATH.
On the other hand, containers providing GPU computing resources on Aethir must stake ATH to provide services, cleverly integrating ATH into the reward and punishment mechanism that maintains system security, while also reducing the circulating supply of tokens in the market.
Simultaneously, after Aethir's DAT buys back tokens, it does not sell them but instead transfers the tokens to the DAT Treasury. These tokens will be used for staking to generate yield or to generate revenue by booking computing resources on the network for enterprise clients, reducing the circulation of ATH in the market.
The expansion of GPU computing power on the Aethir network means that more ATH will be staked. According to Aethir's GPU Dashboard data, over 174 million ATH have been staked so far, with cloud hosts accounting for 54% of that.
Aethir's enterprise-focused economic model design is highly attractive to cloud hosts. Additionally, 80% of the revenue is distributed to cloud hosts, while 20% goes to the Aethir Foundation. Coupled with the continuous reduction in token circulation and the increasing value of token rewards, Aethir can expand the number of cloud hosts on its platform at a tremendous speed.
A large number of cloud hosts joining, increased computing power, and low prices will attract more large clients, thereby creating more revenue and excellent ARR. The reward distribution for cloud hosts increases, GPU supply further expands, and the staking rate and intrinsic value support for ATH from cloud hosts are further enhanced.
Aethir: A Viable Business Case Combining Web3 and AI
AI represents the enhancement of productivity in this era, while Web3 represents the innovation of production relations. However, in the past, the combination of Web3 and AI has been repeatedly disproven, and the market has experienced too many bubble moments. But in the source factory of AI—GPU computing and cloud services—Aethir, through its closed-loop business model, global customer base, positive cash flow, and sound token model, has found a viable business case combining Web3 and AI.
Aethir has genuine cross-border scalability. As the crypto industry becomes more standardized and popularized, although the market's main focus currently remains on areas where Web3 is transforming traditional finance, such as stablecoins and RWA, in the context of the AI-driven era, people will eventually truly realize the innovative significance of Web3 for the AI and computing markets. By then, Aethir, as the leader in decentralized cloud computing power, will be the first to be noticed by capital.
