Aethir achieved Q3 2025 revenue of over $39.8M, representing a ~22 % quarter-over-quarter increase from Q2. The network delivered 340M compute hours (~17 % growth from Q2) and has a cumulative delivery exceeding 1.2B compute hours. Its infrastructure spans over 435,000 GPU containers across 93 countries, with enterprise-grade uptime (~99.31 %).
Why this matters: It shows that Aethir is moving from being a tech play (infrastructure for AI + decentralized GPU) into scaling real, measurable business performance (revenue, compute delivery, global footprint). With global GPU supply constraints and strong demand for AI compute, Aethir’s decentralized model is positioned to capture companies that need scalable GPU access beyond traditional cloud providers.
For the broader DePIN (decentralized physical infrastructure network) / compute-infrastructure sector, it signals that “compute” is becoming a commoditized utility and revenue-driven metrics may become a key differentiator.
Discover Aethir's record-breaking Q3 revenue and learn how its innovative decentralized GPU cloud model is redefining compute scalability for AI innovation. Aethir's Record-Breaking Q3: Stellar Quarterly Growth Redefines Compute Infrastructure
Share Dialog
Albena Kostova-Nikolova
Support dialog
Matrix AI Network has launched M.A.C. (MATRIX AI Capital), an AI-powered investment assistant aimed at the crypto market, positioning itself as a key player in the Crypto x AI ecosystem.
M.A.C. is able to interpret market movements, summarize on-chain activity, and generate clear, actionable insights in plain language. The assistant is built on advanced language models and prediction infrastructure, enabling users to ask questions like “What is the AI-predicted trend for Bitcoin over the next month?” or “Which altcoins are likely to perform well this week?”.
The digital asset manager Galaxy published a detailed report, exploring how decentralized architecture is moving from theoretical to real-world application. It also highlights projects, working to enable permissionless training of foundation models on a global scale and on distributed infrastructure, such as Nous Research, Prime Intellect, Pluralis Research, templar, and Gensyn.
Their motivation is twofold. Practically, they recognize that vast amounts of unused GPUs around the world can be tapped for model training, opening up an otherwise unaffordable process to AI engineers around the world and making open-source AI development a reality. Philosophically, these teams are motivated by the tight control that the leading centralized AI labs wield over one of the most important technological revolutions of our time and the urgent need to create open alternatives.
The report breaks down the core technical challenges of decentralized training, reviews the leading projects' architectures, and describes how they handle distribution of compute, training orchestration, and incentivization of participants. Moreover, it offers an outlook and identifies risks: decentralized training has made real progress, but there remains a gap compared to large centralized labs. Regulatory risks, monetization/distribution hurdles, and maintaining performance parity are emphasized.
However, the report signals that training large AI models is no longer exclusively within the domain of massive centralized entities; decentralized networks could democratize model-development.
If successful, this could reshape the economics of AI: lower cost of entry, open contributions, new incentive models, and potentially more resilience and diversity in model architectures and ownership.
It aligns with broader trends in crypto and compute infrastructure, leveraging global compute resource sharing, token-based incentives, and permissionless networks.
Astra Nova, an AI-crypto project that bills itself as "the first AI Entertainment Ecosystem from the Kingdom of Saudi Arabia", has raised $48.3M to expand its entertainment ecosystem. The funding comprises about $41.6M from a recent strategic round, building on earlier investments from firms such as Outlier Ventures and family offices / institutional investors in the Middle East (Saudi Arabia, UAE, Bahrain).
Astra Nova’s product focus includes no-code creator tools that enable users to launch blockchain-based entertainment experiences: e.g., “TokenPlay AI” (a no-code launchpad), “NovaToon” (a Web3 webtoon platform) and “Deviants: Fight Club” (a Telegram-based PvP game using the firm’s $RVV token). The company aims to expand into new markets across the Middle East, Europe and Asia, and to prepare for regulatory compliance; its strategic partners include NVIDIA Inception, Alibaba Cloud, and NEOM.
Shortly after that, the project reported a loss of ~$10M and blamed it on a third-party market-maker, which began dumping its $RVV token supply, after being hacked. The incident involved about 860M of the $RVV tokens (~8.6% of supply) being converted into $USDT, causing the token price to fall from ~$0.028 to ~$0.007 (decline of ~50%).
Astra Nova said its core smart‐contracts were unaffected, cut ties with the compromised market-maker, and pledged a token buy-back to stabilize the ecosystem. It also offered a 10% bounty if the stolen funds are returned.
The event raises questions: blockchain analysts noted the attacker converting tokens to USDT and transferring to exchanges directly, which may be an odd hacker strategy.
Orochi Network has raised an additional $8M, bringing its total funding to $20M, to support its mission of building a “verifiable data infrastructure”. The company's backers include Ethereum Foundation, MEXC Ventures, Plutus VC, Anti Capital, Presto, and others.
Orochi Network's services aimed at institutional use, especially for tokenizing real-world assets (RWA) and ensuring data integrity under tighter regulation (e.g., the EU’s DORA). It's launched a product, called zkDatabase (mainnet), which reportedly has processed 700+ million proofs and counts over 10,000 downloads plus ~120 enterprise/partner engagements so far. The company is also positioning itself in Vietnam and Southeast Asia as an infrastructure partner and advisor on RWA regulatory sandboxes, signaling both regional expansion and technology-policy integration.
Morpheus' David A. Johnston and Recall's Carson Farmer came together for a talk on decentralized intelligence, AI agents, and the future of the open AI ecosystem.
Johnson, a veteran in the Web3 space, reaffirms the old mantra and brings it into the AI era: just as you own your keys, you should own your AI.
I invite you to listen to the full conversation to get a sense of where the Web3 + AI space is going and what it's striving to achieve.
The Salib Suci Foundation, a K-12 school network in Indonesia, in partnership with Space and Time, is adopting blockchain-based credentialing for approximately 14,000 students across 70 schools. The program is part of a broader initiative by Indomobil Group (one of Indonesia’s large integrated business groups), to build universal verifiable educational records.
By storing the completion records on‐chain, students will be able to present verifiable credentials when applying for further education or employment. The project emphasizes using blockchain to make academic records tamper-proof, portable, and globally verifiable.
Thank you for reading! The next edition is coming tomorrow!
I invite you to subscribe to The Web3 + AI Newsletter to stay in the loop on the hottest dAI developments.
I'm looking forward to connecting with fellow Crypto x AI enthusiasts, so don't hesitate to reach out to me on social media.
Disclaimer: None of this should or could be considered financial advice. You should not take my words for granted, rather, do your own research (DYOR) and share your thoughts to create a fruitful discussion.
113
No comments yet