Domestic inference GPU completes nearly 3 billion in funding; how will the next-generation computing power chips reshape Web3 infrastructure

SerumDegen · 2026-01-22T11:46:18+00:00

Domestic GPU chip company Xiwang announces the completion of nearly 3 billion yuan in strategic financing, with participants including multiple industry and capital institutions. Xiwang emphasizes developing a "more AI-savvy domestic GPU," aiming to reduce the cost of high-end computing chips and support the research and development of next-generation inference GPUs and ecosystem co-construction. This will impact the application of AI models on blockchain and the cost structure of computing infrastructure.

SerumDegen

2026-01-22 11:46:18

Abstract generation in progress

【BlockBeats】Domestic GPU chip company reaches a new financing milestone. Emerging chip company Xiwang announced the completion of nearly 3 billion yuan in strategic financing, with participating investors including industry capital such as Sany Group Huaxu Fund, Fanxi Intelligent, Hangzhou Data Group, as well as top VC/PE institutions like IDG Capital, GGV Capital, and China International Capital Corporation. Additionally, state-owned background forces such as Chengtong Mixed Ownership Reform Fund and Hangzhou Jintou also support the company.

The company was established at the end of 2024, spun off independently from SenseTime’s large chip division. Its core positioning is quite interesting—emphasizing “a domestically produced GPU that understands AI better” and “top-tier products that are affordable and practical.” Compared to the current market situation where high-end computing chips are often exorbitantly priced, this direction hits the pain point directly.

The funds raised will mainly be invested in three areas: next-generation inference GPU development, large-scale mass production, and ecosystem co-construction. From a Web3 perspective, breakthroughs in inference chip performance and cost reduction are directly related to the feasibility of on-chain AI model applications, as well as affecting node operation and on-chain computing economic benefits. As more domestically produced alternatives enter the market, the cost structure of the entire Web3 computing infrastructure may undergo adjustments.

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

10 Likes

Reward
10
2
Repost
Share

Comment

0/400

AirdropHunter

· 01-22 12:12

Wow, 3 billion invested, finally someone cares about cost-effectiveness.

View OriginalReply0

Layer2Observer

· 01-22 12:03

Interestingly, the financing team composition—industry capital + leading VC + state-owned assets—indicates that everyone is optimistic about the logic of domestically produced GPU replacement. But one point needs clarification: a company founded at the end of 2024 raising 3 billion yuan, the speed of financing is indeed fast, and having SenseTime as an endorsement adds credibility. From an engineering perspective, inference chips and training chips are two different things, and the cost reduction potential is indeed greater. The key still depends on actual benchmark data and ecosystem compatibility; theoretically, both are feasible, but only real-world application counts. Regarding Web3, I remain cautious. Unless they are specifically working on on-chain computation optimization, the practical scenarios still need to be explored.

View OriginalReply0