self-developed

According to a report by Central News Agency citing Reuters, Microsoft announced the launch of its third-generation self-developed AI chip, Maia 300, with plans to deploy over 1 million units next year to support internal AI services. The small chip responsible for key computing functions uses TSMC's 2-nanometer process. The Maia 300 can handle code writing and perform complex business tasks on behalf of users. The previous generation of chips has been used to run Microsoft's internal AI models and the OpenAI models provided through Microsoft Azure cloud computing services.Microsoft is taking a different technological approach compared to NVIDIA. The NVIDIA AI accelerator uses a large computing chip surrounded by high-bandwidth memory chips; the Maia 300 core processing unit is composed of small chips, with the small chips responsible for key computing functions using TSMC's 2-nanometer process, while other small chips use the lower-cost 3-nanometer process to handle data transmission out of the processor and transfer data to other locations within the system.The Maia 300 will be installed in cabinets, with each cabinet accommodating 96 chips and consuming about 167 kW of power. In contrast, NVIDIA's latest GB300 AI server cabinet is configured with 72 GPUs and consumes about 142 kW of power. Microsoft stated that it plans to deploy the Maia 300 on a large scale starting next year, reaching several gigawatts. Microsoft previously indicated that it expects its capital expenditures to exceed $175 billion this year.

According to Business Insider, Thomson Reuters launched its first self-developed AI model, Thomson-1, on Monday. This model is based on Snowdon and constructed through "re-alignment" of Alibaba's open-source Qwen model. Open-source means anyone can download and modify the model for free. Thomson-1 will take over some tasks previously handled by Anthropic's Claude, but it is not intended to completely replace collaboration with Anthropic and other labs, initially focusing on the company's areas of expertise, starting with document review.This move aims to address the high AI costs brought by models like Claude and OpenAI Codex, and reflects the use of cheaper Chinese open-source AI. CTO Joel Hron stated that the main reason is to better leverage Thomson Reuters' own expertise and control costs. The company expanded its collaboration with Anthropic in May this year for the AI legal assistant CoCounsel, which still primarily relies on Claude. Hron said, "Our main goal is to gradually make Thomson the model that drives more and more capabilities for CoCounsel."Thomson Reuters, in collaboration with a team from Imperial College London, spent months transforming Qwen to build Snowdon and ensured it is "ethically and politically bias-processed and safe to use." Hron pointed out that having a proprietary model allows for development based on its own intellectual property rather than continuously paying external AI companies, comparing it to renting versus buying a house: renting provides shelter but does not accumulate long-term equity.

According to the Wall Street Journal, Wells Fargo plans to launch Tokenized Deposits services aimed at corporate clients, further exploring the application of blockchain technology in the traditional financial system.Wells Fargo stated that the tokenized deposit system will operate on its self-developed blockchain network and support interoperability with other blockchain networks.Tokenized deposits represent bank deposits in a digital form on the blockchain, allowing for faster and more transparent transfers of funds on-chain, while still being supported by the bank's deposit system.This move signifies that large banks are accelerating the construction of blockchain infrastructure and exploring on-chain financial applications in scenarios such as corporate payments, fund management, and inter-institutional settlements.

According to a report by The Information, multiple informed sources revealed that China's leading AI company Zhiyu is considering designing its own AI chips. It is reported that as the demand for its GLM series large models has rapidly surged, coupled with increasing constraints on computing resources, Zhiyu has recently had preliminary contacts with several domestic chip design companies to explore the possibility of collaborating on the development of custom AI processors.The report pointed out that for large AI model developers, the layout of self-developed chips aims to achieve software and hardware collaborative optimization, improve computing power utilization efficiency, and reduce dependence on external GPU suppliers. In addition to Zhiyu, another domestic large model company, DeepSeek, has also been reported to be simultaneously advancing its self-developed chip project focused on inference scenarios. Globally, top AI developers such as OpenAI (which has announced its self-developed chip Jalapeño) and Anthropic are actively laying out custom chips, and the AI industry is accelerating towards a dual-line layout phase of "model + chip" software and hardware integration.

In response to the pain points of trading US stock tokens on crypto platforms, Xie Jiayin recently expressed his views, stating that after Bitget entered UEX in the third quarter of last year, its market share of US stock tokens once reached 90%. However, the early products had shortcomings in liquidity, dividend distribution, and stock split experience.To address this, Bitget is about to launch its self-developed product Reality, aimed at solving the above issues. According to the introduction, this product has the following three major features:Direct connection to broker liquidity: Provides a trading experience with no price difference and high liquidity;Supports dividend distribution: Dividends can be directly distributed to user accounts;Enhances capital utilization: US stock tokens can be used as margin and support participation in strategy trading and lending wealth management, etc.In addition, Xie Jiayin revealed that rToken will officially launch on Bitget in early June.

Paradigm officially announced the open-sourcing of Centaur, a self-hosted AI Agent Runtime developed jointly by Paradigm and Tempo that supports multi-user collaboration and secure operation. Paradigm stated that since January of this year, the fund has been using Centaur internally, which has completely transformed the way the team works, covering various task scenarios such as investment, engineering, design, recruitment, events, and customer service.Centaur can be understood as a "shared Agent" that can call tools, run continuously for hours or even days, continue working after a system restart, and perform operations using real credentials without ever touching the raw secrets. Users can communicate with it via Slack or call it through an API.

ChainCatcher news, according to The Block, Akash Network founder Greg Osuri stated that the project will abandon its self-developed Cosmos SDK chain and migrate to a new network that can "secure Akash," aiming to maintain IBC compatibility.The team has not released a timeline and will publicly evaluate chain candidates, emphasizing strong security, high-quality community, deep liquidity, and growth potential, while retaining integrations with Osmosis, Keplr, and others. Osuri mentioned that the staking mechanism will "evolve to be better" and will not be limited to consensus security, with an RFP to be released in a few weeks; Solana is referred to as a "strong candidate." Akash introduced Nvidia GPUs in 2023 Mainnet 6, providing decentralized computing power at a lower cost.

ChainCatcher news, according to The Block, Franklin Templeton is expanding its Benji technology platform to the BNB Chain, aiming to broaden investors' access to tokenized investment products.The Benji platform is a tokenization platform independently developed by Franklin Templeton, designed to support the trading, management, and operation of token-based investment products. The platform was used in 2021 to launch the world's first on-chain registered U.S. mutual fund and currently supports multiple tokenized products for the company, serving both retail and institutional clients.

ChainCatcher news, according to Jinshi reports, Ant Group's Vice President and President of Ant Digital Technology's Blockchain Business, Bian Zhuoqun, revealed today that the company's self-developed public chain is scheduled to officially launch by the end of September 2025, and this public chain will be able to handle real transactions.On April 30 of this year, Ant Digital Technology released the Layer2 blockchain Jovay, clearly stating that Jovay is a compliance-grade blockchain and will not issue tokens. Bian Zhuoqun stated that Ant Digital Technology is committed to becoming a gateway connecting Web2 and Web3, providing more trusted assets for Web3 while introducing more compliant funds for Web2. She cited data predicting that the global tokenization market size is expected to reach $16 trillion by 2030.Currently, Ant Digital Technology has been expanding its RWA asset landscape in fields such as new energy, computing power, and finance.

ChainCatcher news, according to Jinshi Data citing the announcement from Jingwei Tian Di Hong Kong Stock Exchange, the company announced its entry into the cryptocurrency payment sector and will launch its first mobile application "Fopay" on July 21, 2025, providing a one-stop payment platform.Fopay is developed based on the concept of stablecoin cryptocurrency payments and currently offers stablecoin custody and prepaid card payment functions through several licensed partners. The board believes that the launch of Fopay and the new business segment can explore more business opportunities for the company and be beneficial to all shareholders.

ChainCatcher news, according to Jin10 data reports, Ant Group's technology arm has opened four self-developed technologies for the Hong Kong market------Layer2 network, large model development tools, "Blockchain + IoT" trusted architecture, and institutional-level Web3 wallet technology, providing full-stack technical services for building a global digital asset innovation center in Hong Kong. Previously, Ant Group's technology arm joined the Hong Kong Monetary Authority's Ensemble sandbox as a core member and announced that its overseas headquarters will be located in Hong Kong.

ChainCatcher message, Xterio's self-developed game Age of Dino, which is invested by Binance, has completed the launch of its core PFP--Dinosty today. Within 12 hours, over 20,000 addresses participated, achieving a 335% oversubscription, with a TVL exceeding 10,000 ETH.It is reported that NFT holders will receive airdrops of the game's dual tokens, with the highest gold mining benefits, and a soft staking plan will disclose details this week.

ChainCatcher news, CoinDesk reports that Block, the fintech payment company founded by Twitter co-founder Jack Dorsey, has recently purchased a large number of Bitcoin mining chips from Intel, accelerating its plans to enter the mining hardware market.Previously, Intel announced in February 2022 that it would cease production of application-specific integrated circuit (ASIC) chips for Bitcoin mining, with the final production date set for April 2024. Block plans to complete the design of its 5nm chips for Bitcoin mining this quarter and manufacture machines based on this. This procurement means the team can focus on the 3nm design.Block's hardware lead Thomas Templeton stated that Block's first-party products will be launched early next year. When asked about the number of chips they purchased, he said it was enough to cover the time needed to design and produce their own 3nm chips. (source link)

ChainCatcher news, the one-stop Web3.0 security development environment and middleware engine Hamster has announced comprehensive support and compatibility for the Move-based Sui network. As of now, Hamster has successfully completed the automated pipeline invocation of Sui contract templates based on the Move language and has opened a one-stop node deployment RPC service for the Sui network. With this, developers within the Sui ecosystem can modularize, toolize, and streamline their development processes, allowing for quick onboarding and integration into the ecosystem.It is reported that as a one-stop developer platform infrastructure, Hamster's comprehensive integration and compatibility with the Sui network will significantly lower the development threshold for developers entering the Sui network and improve development efficiency. In addition, multiple Move language contract templates based on the Sui ecosystem's DeFi, NFT, and other tracks are currently being iterated and updated. (Source link)

ChainCatcher news, metaverse chip manufacturer Nano Labs recently launched its self-developed A series iPollo computing device, which is equipped with high-performance graphics cards featuring high video memory and multi-core CPUs, aimed at supporting artificial intelligence generated content (AIGC) and zero-knowledge proofs (ZKP) computations. Zero-knowledge proofs play an important role in the crypto space and Web 3.0, and are highly anticipated in the expansion of Ethereum and the development of the blockchain industry. Marvin Kong, Senior Vice President of Nano Labs, stated that this product can support the computational needs of Web3, the metaverse, and other fields.Nano Labs successfully went public on Nasdaq last July, and its iPollo and iPolloverse fully support the creation of the underlying platform for the metaverse. (Source link)