Aethir · 103

Decentralized AI data network Perceptron Network completes $6.5 million strategic financing, Selini Capital and others participate

Comparing news, the decentralized AI data network Perceptron Network completed $6.5 million in strategic financing, with investors including Sigma Capital, Selini Capital, QCP Capital, P2 Ventures, CoinDCX Ventures, Momentum6, DeFi Capital, Walrus Foundation, Aethir, Colosseum, GuruDev Capital, Tempo Finance, NewTribe Capital, Digital Consensus Fund, and CodeCraft Capital. The funding round will support the launch of Perceptron's data exploration platform, expand its contributor tools and rewards infrastructure, and scale the network to its goal of 5 million nodes. Perceptron aims to consolidate global data collection into a single contributor network, enabling AI companies and ordinary contributors to acquire, validate, and monetize high-quality data sets. Perceptron is currently online and has over 700,000 nodes.

21d ago#financing

Perceptron Raises $6.5 Million to Accelerate Construction of Decentralized AI Data Infrastructure

Comparing news, decentralized AI data network Perceptron announced the completion of a $6.5 million strategic financing. Participants include a number of Web3 investment institutions, trading firms, ecosystem partners and infrastructure companies, including Sigma Capital, Selini Capital, QCP Capital, P2 Ventures, CoinDCX Ventures, Momentum6, DeFi Capital, Walrus Foundation, Aethir, Colosseum, GuruDev Capital, Tempo Finance, NewTribe Capital, Digital Consensus Fund, and CodeCraft Capital. The funding round will be used to launch Perceptron's data task platform, expand contributor tools and reward infrastructure, and drive the network's expansion towards the target of 5 million nodes. Perceptron aims to make real-world data acquisition fast and convenient enough so that AI companies no longer need to rely on centralized data capture infrastructure. The network consolidates global data collection into a unified network of idle bandwidth, unique data sets, and domain expertise. Through this unified form, both individual contributors and AI companies can generate complete and validated data sets, reducing the time from requesting the data to delivering the data set to a few days. Perceptron is in between two models: one end is a centralized data scraping tool with limited capabilities, and the other end is a closed data collaboration that is difficult for most AI startups to qualify for. Perceptron provides data that AI companies and ordinary contributors can directly use. Through Perceptron, contributors own their own data and revenue, which can be monetized or extracted at any time without relying on a third party. Perceptron is geared towards continuous contribution and continuous iteration rather than one-time data capture. Perceptron will focus on launching a data mission platform in the near future, and it is expected that more announcements will be made in the next quarter. Over the long term, Perceptron is moving towards a fully integrated network of 5 million nodes, so that everything AI companies need is centralized on the same platform to achieve a seamless connection between demand and supply.

22d ago

Aethir launches self-developed open source LLM API platform Aethir Mesh, DeepSeek V4, Kimi K2.6 and other top models run all inference on Aethir GPUs

According to official news, according to official news, Aethir officially released Aethir Mesh — a self-developed open source large model API platform running on Aethir's decentralized GPU infrastructure. Currently, it provides unified API endpoints directly connected to mainstream open source models such as DeepSeek V4, Kimi K2.6, GLM-5.1, MiniMax-M2.5, and Qwen3.6-27B. Unlike aggregated routing solutions such as OpenRouter, Aethir Mesh uses Aethir's own computing power to directly provide inference services, no resale premiums, and no third-party routes. Aethir Claw users can directly write the mesh API key into the agent configuration to achieve a full-stack closed loop from VPS hosting to model inference, and the inference data does not leave the Aethir ecosystem throughout the process. The platform is now online: https://agent.aethir.com/twitterCNmesh

79d ago

Aethir officially launched the delivery of computing power for Axe Compute's $260 million B300 cluster, and the first payment of $43 million has been paid

Comparatively, Axe Compute (NASDAQ: AGPU) announced that the underlying infrastructure provider for the $260 million enterprise contract is Aethir, a globally distributed GPU infrastructure. According to reports, Axe Compute has confirmed receipt of the first payment of 43 million US dollars under the contract, and Aethir has officially launched the computing power delivery process for 2,304 NVIDIA B300 GPU clusters. The target launch time is the third quarter of 2026. The contract has a 36-month non-negotiable structure and is expected to generate approximately $21 million in quarterly revenue once it goes live.

84d ago

Aethir Infrastructure signs $260 million enterprise-wide contract from Axe Compute

According to news, Nasdaq-listed company Axe Compute Inc. (NASDAQ: AGPU) announced today that it has signed a 36-month AI infrastructure contract with enterprise customers with a total contract value of about US$260 million. The deployment includes 2,304 NVIDIA B300 GPU dedicated clusters and high-speed AI storage systems, which is expected to be officially launched in the US Tier 3 data center in the third quarter of 2026. This is Axe Compute's largest single enterprise contract since its inception. Axe Compute's underlying computing power infrastructure is supported by the Aethir decentralized GPU cloud. The contract includes exclusive 4.8 MW of N+1 redundant electricity, the customer specifies the deployment site and service standards, and uses structured payment arrangements. Daniel Wang, co-founder and CEO of Aethir, said the contract was the result of months of deep collaboration between the two teams: “The enterprise-grade workload supported by 2,304 NVIDIA B300 GPUs is exactly the scale we were aiming for when we first built Aethir's computing power infrastructure. Today's announcement proves that decentralized GPU infrastructure is capable of taking on AI workloads from the world's top enterprises.” After the news was released, the stock price of Axe Compute (NASDAQ: AGPU) fluctuated sharply. The highest daily increase was over 166%, hitting $14.47. The daily trading volume increased more than 100 times compared to the three-month average daily level.

121d ago

Since Drift Protocol was hacked, at least a dozen cryptographic entities have been attacked

Comparatively, according to Cointelegraph, at least 12 DeFi protocols and cryptocurrency companies have been attacked in just over two weeks since Drift Protocol was attacked for $280 million on April 1. Since early April, attacks on crypto protocols or companies have included CoW Swap, Hyperbridge, Bybit, Dango, Silo Finance, BSC TMM, Aethir, MONA, Zerion, and most recently Rhea Finance and Grinex exchanges. Among them, DeFi protocol Rhea Finance reported that the attackers “used a flaw in Rhea's margin trading function to perform a coordinated fund pool manipulation attack”, affecting the Rhea Lend smart contract. According to blockchain security firm CertiK, the stolen funds were around $7.6 million.

127d ago

Aethir successfully isolated cross-chain contract attacks, and the affected amount was kept within $90,000

Comparing news, decentralized cloud infrastructure provider Aethir tweeted that it has detected and successfully isolated malicious attacks against ATH cross-chain contracts (connecting Ethereum to other chains). Currently, all affected contracts have been disconnected, and the vulnerability has been brought under control; the total supply of ATH on the Ethereum main network has not been affected; the ETH-ARB cross-chain bridge on Squid has not been affected; the actual amount of user impact is kept within $90,000; the team will announce the full compensation plan next week. Aethir is currently working with several law enforcement agencies and exchanges (Binance, Upbit, Bithumb, HTX, etc.) to freeze funds and track down attackers. Additionally, the team has announced the addresses of affected contracts, reminding users not to interact with them for the time being.

134d ago

Axe Compute released its first NASDAQ annual earnings report, and decentralized computing power entered the eyes of institutional investors for the first time

Comparatively, Axe Compute (NASDAQ: AGPU), a NASDAQ listed company built on Aethir's decentralized GPU network, released its FY2025 annual financial report, becoming the first company to disclose the progress of decentralized computing power business to institutional investors in the form of annual results of listed companies. Financial reports show that the total number of contracts executed in the past 30 days is about 12 million US dollars, the estimated monthly revenue for Q2 is 835,000 US dollars, and there are more than 20 enterprise customers. After the financial report was released, AGPU had the highest increase of +152% in a single day. The daily trading volume exceeded the 20-day average of 900 times, and the computing power deployment of all enterprises with a total daily transaction volume of more than 300 million US dollars was delivered by Aethir Network, covering 94 countries, more than 200 nodes, over 435,000 GPU containers, and delivering a total of more than 1.8 billion hashrate hours.

142d ago
The copycat won't die; it will only become more and more like the “US stock market”...

The copycat won't die; it will only become more and more like the “US stock market”...

Author: TVBee Original title: Altcoins/VC coins will not die out; they will only tend to become US stocks... The contradiction between the copycat season and the redundancy of VC coins actually happened around May to August 2025. BTC is rising, and at the same time, its market share is declining. The market share of altcoins is rising at this stage. BTC is rising, while the total market value of copycats is growing even more. This is the copycat season. The contradiction of VC coin's redundant copycat season is that the total market value is growing, yet the market has no perception of the copycat season. This is because the total market value of altcoins is growing, but there are so many altcoins that the vast majority of individual altcoins have not risen significantly. So the key problem with the 2025 counterfeit season is — VC coin redundancy. So the question is, why will there be so many VC coins in 2025? The normal route of misalignment in the VC coin 1 and 2 market liquidity should be for VC to invest first, then enter the secondary market after the token TGE. Facing greater liquidity in the secondary market, the price rises. However, in this round of the market, there was a liquidity misalignment between the primary market and the secondary market. ◆ The reason for VC coin redundancy is excessive liquidity from 2021 to 2022. There was a large amount of VC investment, investment amount, and number of projects in the primary market. Financing in other time periods is generally only concentrated within 1 to 2 months, but from 2021 to 2022, there is a lot of financing and the duration is long. We can think that this is financing redundancy due to excessive macro liquidity, which is the root cause of VC coin redundancy in this round of the market. ◆The reason for the poor performance of VC coin prices is the lack of liquidity in the secondary market. However, whether it is the previous M2 growth rate or the financing data map, it fully explains that market liquidity is declining after these redundant VC coin TGE in 2025. As a result, this kind of time gap between the liquidity in the primary market and the secondary market was created. A large number of VC coins have redundant liquidity in the primary market financing stage, while in the secondary market stage after TGE, there is a lack of liquidity. Of course, there are many reasons for the lack of liquidity in the secondary market, which will be discussed in the next article. Summarizing the most critical issue is that the liquidity misalignment in the VC coin tier 1 and 2 market led to a relatively large number of VC coins during the 2025 copycat season, and the selling power was relatively high. Instead, the buying power was insufficient, and the currency price performance was not ideal. Altcoin/VC coins will not die out. VC coin financing continues with a more cautious attitude and is still the previous VC financing plan. As can be seen, there are still quite a few projects receiving financing in 2025. Including the total amount of financing and the number of projects, it is far lower than in 2021. On the one hand, it may be that macro-liquidity is relatively unrelaxed; on the other hand, it also shows the cautious attitude of VCs in investment activities. One thing to note, however, is that in 2025, an average of about 75 projects were funded each month. The amount and volume of VC financing in 2025 was higher than in 2017-2018. This shows one thing: after we ignored VC coin financing redundancy in 2021, altcoin/VC coin financing is still improving. After this round of the bear market is over, after VC coins are TGE again, there will be no redundancy of as many new coins as in 2025. Unless macro-liquidity is still not relaxed enough, the new round of counterfeit market will probably be one of two situations between the craziness of 2021 and the cold of 2025. Web3 finance still holds a scale of nearly 100 billion US dollars. First, in the financial sector, a large number of TRADFI institutions have joined Web3. There is no need to even give an example in this area anymore. Currently, the TVL of the entire network is still 92,831 billion US dollars, which is roughly equivalent to the level of March to April 2021. Technology companies and Web3 continue to integrate ✦ Examples of tech giants participating in Web3 are in the non-financial sector, and tech giants are also participating in Web3. For example, Google is deeply involved in the Hedera ecosystem, is a network node of Hedera, and is also a member of its management committee. Hedera's management committee members also include IBM, Boeing, and Nomura Securities. Google has also integrated Hedera's ledger data into BigQuery, and developers can use SQL to query Hedera blockchain data. (Blockchain is not a relational database and does not support direct use of SQL. Generally, SQL can only be used after indexing). Google is also one of the Midnight (Cardano sidechain with privacy protection) nodes, and provides developers with development tools to quickly deploy Midnight nodes using Google Cloud. Nvidia recently boosted its AI...

143d agoLuxurytracy#Market topics

Aethir Announces Launch of Physical-Level Isolated AI Agent Hosting Platform Aethir Claw Alpha Closed Beta

In comparison, according to official sources, Aethir announced the launch of Aethir Claw Alpha, a physical-level isolated AI proxy hosting platform. According to reports, Aethir Claw completely broke the traditional hosting service provider's “rent a third-party cloud and then resell” model, and directly used Aethir's own distributed computing power resources worth more than 400 million US dollars. With Aethir Claw, users can not only obtain a completely isolated VPS environment, but also choose a “zero provider access” mode to ensure absolute root-level data sovereignty over the AI agent's API keys, conversation records, and execution actions. At the same time, the platform is pre-installed with OpenClaw and is deeply customized for business scenarios in the next stage. It is planned to build in cryptographic native skills such as token tracking, on-chain alerting, and giant whale monitoring, and support cryptocurrencies such as ATH and USDT and fiat payments. At the same time, a “Model as a Service” (MaaS) layer will be introduced in the next phase to introduce open source big model reasoning and graphic and video generation functions running on Aethir GPUs to fully enter the agent-based AI (Agentic AI) application layer market. Aethir provides enterprise-level customers with Kubernetes GPU clusters to automatically manage large-scale server clusters. Across 94 countries and regions, 435,000 GPU containers provide extremely low latency computing power access to more than 200 nodes, reducing overall computing power costs by 60% to 80%.

150d ago