广告
加载中

商汤分拆芯片企业曦望完成20亿融资 估值达200亿

亿邦AI 2026-09-01 16:55
亿邦AI 2026/09/01 16:55

邦小白快读

EN
全文速览

本文核心信息为商汤分拆出的AI芯片企业曦望完成20亿元融资,投后估值达到200亿元,相关干货信息如下:

1. 曦望源自商汤2024年底推出的1+X战略,商汤此前在芯片研发领域累计投入超过11亿元,独立后主打大模型推理芯片研发;核心团队规模近200人,联席CEO分别是前百度创始成员王湛、前AMD及昆仑芯芯片老将王勇,核心成员多来自头部科技和芯片企业。

2. 曦望已经量产两款芯片,其中S1云边视觉推理专用芯片累计出货超万片,NPU IP已经授权给索尼、小米等企业,2024年量产S2大模型推理GPGPU,兼容CUDA架构,下一代S3芯片依然瞄准推理计算市场。

3. 当前AI算力需求向推理侧倾斜,2026年推理算力占全部AI算力的三分之二,市场规模突破3000亿元,本次融资将用于曦望的产品研发和商业化布局。

本文给布局AI相关业务的品牌商提供了清晰的产业趋势和商业参考,核心干货如下:

1. 产业趋势层面,当前AI产业需求已经发生变化,算力需求从训练侧逐步向推理侧倾斜,2026年推理算力占全部AI计算能力的三分之二,整体市场空间突破3000亿元,大模型部署环节的国产替代需求明确,国内企业都在加大相关领域投入,赛道增长空间十分明确。

2. 产品研发与合作参考层面,头部AI企业商汤选择将自研芯片业务分拆独立运营,通过市场化融资加快发展,这种模式给品牌布局AI底层技术研发提供了参考;当下国产推理芯片已经进入量产落地阶段,已经有产品进入索尼、小米等头部品牌供应链,布局AI应用的品牌可以对接本土芯片厂商,降低供应链成本,推进国产替代方案落地。

本文给AI相关领域的卖家传递了最新的行业机会与风向,核心干货如下:

1. 增长市场层面,当前AI推理赛道已经成长为千亿级规模市场,2026年市场空间突破3000亿元,占整体AI算力的三分之二,需求结构变化带来了新的增长机会,国产替代的大背景下,本土推理芯片企业获得了很大的发展空间,属于当前的蓝海增长赛道。

2. 机会与合作提示层面,商汤分拆的曦望已经完成两款芯片量产,拿到20亿元融资后将加速拓展研发和商业化,上下游卖家可以对接这类快速成长的国产芯片企业,挖掘合作机会:比如给下游AI应用厂商对接芯片供应,或者围绕国产芯片开发配套的应用解决方案,都有不错的市场机会。

3. 行业动向显示大模型落地对国产推理芯片需求旺盛,相关领域的创业和合作机会较多,卖家可提前布局相关方向。

本文给芯片相关制造工厂以及传统制造工厂都提供了有用的参考信息,核心干货如下:

1. 产品生产需求与商业机会层面,当前大模型推理芯片市场需求爆发,2026年推理算力占整体AI算力的三分之二,市场规模突破3000亿元,国产替代需求强烈,曦望作为头部成长型芯片企业,拿到20亿元融资后会扩大产品研发和量产,上游芯片代工、零部件生产工厂可以对接这类企业,拓展订单来源,挖掘新的增长机会。

2. 曦望已经实现两款芯片量产,S1累计出货超万片,NPU IP已经进入索尼、小米等头部企业供应链,接下来商业化布局会加快,对上游生产端的需求会持续提升。

3. 对于想要推进数字化转型的传统工厂,国产推理芯片的快速发展会带来更多低成本的AI解决方案,有利于工厂降低数字化转型的投入成本,后续可以关注相关领域的技术落地。

本文给AI产业链相关服务商提供了明确的行业趋势和客户需求信息,核心干货如下:

1. 行业发展趋势层面,当前AI算力需求结构已经发生根本性变化,推理侧需求占比已经超过训练侧,2026年推理算力占全部AI算力的三分之二,整体市场规模突破3000亿元,国产替代是明确的发展方向,大量本土AI芯片企业正在快速崛起,给产业链服务商带来了很多新的客户需求。

2. 客户痛点层面,成长中的国产芯片企业拿到融资后,快速拓展研发和商业化,需要招聘高端人才、对接供应链资源、拓展下游客户、品牌营销等多方面的服务支持,而下游AI应用企业也需要对接靠谱的国产芯片供应商,存在明确的对接需求。

3. 服务商可以围绕这类快速成长的国产芯片企业,推出针对性的服务产品,拓展自身的客户群体,抓住国产替代带来的行业增长红利。

本文给AI产业平台、算力服务平台等平台商提供了清晰的布局方向,核心干货如下:

1. 市场需求层面,当前AI推理芯片市场已经达到千亿级规模,2026年推理算力占整体AI算力的三分之二,市场需求旺盛,国产推理芯片企业正在快速崛起,平台商可以针对性调整自身的业务布局,围绕国产推理芯片产业链打造配套服务,满足上下游企业的对接需求。

2. 招商与运营层面,曦望作为刚刚完成20亿元融资的头部高成长AI推理芯片企业,接下来会大力拓展商业化布局,平台商可以将这类企业作为重点招商对象,引入平台后可以提升平台自身的产业影响力和资源厚度,吸引更多上下游企业入驻。

3. 风险规避层面,国产替代是AI芯片领域的明确发展风向,平台商提前布局国产芯片相关的产业服务,可以规避依赖海外芯片带来的政策和供应链风险,抓住赛道风口实现自身增长。

本文给AI产业、半导体产业的研究者提供了最新的产业动向和研究样本,核心干货如下:

1. 产业新动向层面,当前全球AI产业发展进入新的阶段,算力需求结构发生重大变化,推理算力的占比已经提升到三分之二,2026年整体市场规模突破3000亿元,推理芯片成为当前AI领域最热门的赛道之一,国内正在加速推进大模型部署环节的国产替代,头部AI企业纷纷加大相关领域投入,产业格局正在发生变化。

2. 商业模式新样本:头部AI企业商汤通过1+X战略,将积累多年的芯片业务分拆为独立公司曦望,前期累计投入超11亿元研发,分拆后独立融资,估值在短时间内接近翻倍,这种内部孵化后分拆独立运营融资的模式,给AI企业硬科技业务孵化提供了新的研究样本。

3. 本文披露了多个权威行业数据和企业运营数据,包括融资额、估值、出货量、市场占比、市场规模等,可供产业研究作为基础数据使用。

返回默认

声明:快读内容全程由AI生成,请注意甄别信息。如您发现问题,请发送邮件至 run@ebrun.com 。

我是 品牌商 卖家 工厂 服务商 平台商 研究者 帮我再读一遍。

Quick Summary

This article covers the core news that Xiwang, the AI chip spin-off from SenseTime, has completed a RMB 2 billion financing round, pushing its post-investment valuation to RMB 20 billion. Key takeaways are as follows:

1. Xiwang originated from SenseTime's "1+X" strategy launched at the end of 2024. SenseTime has invested a total of over RMB 1.1 billion in chip R&D to date. After becoming independent, the company focuses on large model inference chips. It has a core team of nearly 200 people, with co-CEOs Wang Zhan, a founding member of Baidu, and Wang Yong, a veteran chip engineer previously with AMD and Kunlunxin. Most core team members hail from leading technology and chip companies.

2. Xiwang has already mass-produced two chip products. Its S1, a dedicated edge-cloud visual inference chip, has shipped over 10,000 units, and its NPU IP has been licensed to companies including Sony and Xiaomi. The S2, a GPGPU for large model inference, will enter mass production in 2024 and is compatible with the CUDA architecture. Its next-generation S3 chip will also target the inference computing market.

3. AI computing demand is increasingly shifting toward inference. By 2026, inference computing will account for two-thirds of all AI computing capacity, with the market size exceeding RMB 300 billion. The new financing will be used to advance Xiwang's product R&D and commercial expansion.

This article provides clear industry trend insights and business references for brands developing AI-related businesses, with key takeaways as follows:

1. In terms of industry trends, AI demand has shifted, with computing demand gradually moving from model training to inference. By 2026, inference computing will account for two-thirds of all AI computing capacity, and the overall market size will exceed RMB 300 billion. Domestic substitution demand for large model deployment is clear, and Chinese companies are ramping up investment in this space, making the growth trajectory of this track highly certain.

2. For product R&D and cooperation reference: Leading AI firm SenseTime has chosen to spin off its self-developed chip business into an independent entity to accelerate development via market financing. This model offers a reference for brands looking to build their own foundational AI R&D capabilities. At present, domestic inference chips have entered mass production and deployment, with products already adopted into the supply chains of leading brands including Sony and Xiaomi. Brands developing AI applications can partner with local chip makers to lower supply chain costs and roll out domestic substitution solutions.

This article shares the latest industry opportunities and trends for sellers operating in AI-related fields, with key takeaways as follows:

1. In terms of growth markets: The AI inference track is already a hundred-billion-dollar scale market, and it will exceed RMB 300 billion by 2026, accounting for two-thirds of all AI computing. Shifting demand structure has created new growth opportunities, and against the backdrop of domestic substitution, local inference chip companies have gained significant room for development, making this a current blue ocean growth track.

2. For opportunity and cooperation tips: SenseTime spin-off Xiwang has already mass-produced two chips, and its new RMB 2 billion financing will accelerate R&D and commercialization. Upstream and downstream sellers can partner with these fast-growing domestic chip companies to explore cooperation opportunities: for example, matching chip supply to downstream AI application developers, or developing supporting application solutions around domestic chips, both of which offer strong market potential.

3. Industry trends show that large model deployment has created robust demand for domestic inference chips, with abundant startup and cooperation opportunities in this space. Sellers can布局相关方向提前 to capture these opportunities.

This article provides useful references for both chip manufacturing facilities and traditional manufacturing plants, with key takeaways as follows:

1. In terms of production demand and business opportunities: Large model inference chip demand is currently booming. By 2026, inference computing will account for two-thirds of all AI computing capacity, with the total market size exceeding RMB 300 billion, and domestic substitution demand is strong. As a leading high-growth chip company that has just secured RMB 2 billion in financing, Xiwang will expand its R&D and mass production capacity. Upstream chip foundries and component manufacturers can partner with such companies to expand their order books and unlock new growth opportunities.

2. Xiwang has already achieved mass production of two chips: the S1 has shipped over 10,000 units, and its NPU IP has entered the supply chains of leading firms including Sony and Xiaomi. It will accelerate commercial expansion next, which will drive sustained growth in demand for upstream production.

3. For traditional factories looking to advance digital transformation, the rapid development of domestic inference chips will bring more low-cost AI solutions, helping to reduce investment costs for transformation. Plants can follow technology deployment in this space going forward.

This article provides clear industry trend and customer demand insights for service providers across the AI industry chain, with key takeaways as follows:

1. In terms of industry development trends: The structure of AI computing demand has fundamentally changed, with inference-side demand already exceeding training-side demand. By 2026, inference computing will account for two-thirds of all AI computing capacity, with the overall market size exceeding RMB 300 billion. Domestic substitution is a clear development direction, and a large number of local AI chip companies are growing rapidly, bringing abundant new customer demand for industry chain service providers.

2. In terms of customer pain points: High-growth domestic chip companies will expand R&D and commercialization rapidly after securing financing, and require a wide range of supporting services including executive recruitment, supply chain matching, downstream customer development, and brand marketing. Meanwhile, downstream AI application companies also need access to reliable domestic chip suppliers, creating clear matching demand.

3. Service providers can develop targeted service offerings for these fast-growing domestic chip companies to expand their customer base and capture the industry growth dividend brought by domestic substitution.

This article provides clear layout directions for platform operators including AI industry platforms and computing service platforms, with key takeaways as follows:

1. In terms of market demand: The AI inference chip market is already worth hundreds of billions of RMB. By 2026, inference computing will account for two-thirds of all AI computing capacity, and market demand is robust. Domestic inference chip companies are growing rapidly, so platforms can adjust their business layouts accordingly, build supporting services for the domestic inference chip industry chain, and meet the matching needs of upstream and downstream enterprises.

2. In terms of investment promotion and operation: As a leading high-growth AI inference chip company that has just completed a RMB 2 billion financing round, Xiwang will aggressively expand its commercial footprint. Platforms can target this type of company as a key recruitment target; onboarding such firms can enhance the platform's industry influence and resource depth, attracting more upstream and downstream companies to join the platform.

3. In terms of risk mitigation: Domestic substitution is a clear development trend in the AI chip sector. By building out domestic chip-related industry services in advance, platforms can avoid policy and supply chain risks that come with relying on foreign chips, and capture growth by riding the track's boom.

This article provides the latest industry developments and research samples for researchers focused on the AI and semiconductor industries, with key takeaways as follows:

1. In terms of new industry trends: The global AI industry has entered a new phase of development, with major shifts in computing demand structure. By 2026, inference computing will account for two-thirds of all AI computing capacity, with the overall market size exceeding RMB 300 billion. Inference chips have become one of the hottest tracks in AI today, and China is accelerating domestic substitution for large model deployment. Leading AI companies are ramping up investment in this space, and the industry landscape is shifting.

2. In terms of new business model samples: Leading AI firm SenseTime spun off its years-long accumulated chip business into independent company Xiwang as part of its 1+X strategy. SenseTime invested over RMB 1.1 billion in R&D upfront, and the spin-off raised independent financing, with its valuation nearly doubling in a short period of time. This model of internal incubation followed by spin-off, independent operation and financing offers a new research sample for hard tech business incubation at AI companies.

3. This article discloses multiple authoritative industry and operational data points including financing amount, valuation, shipment volume, market share and total market size, which can be used as foundational data for industry research.

Disclaimer: The "Quick Summary" content is entirely generated by AI. Please exercise discretion when interpreting the information. For issues or corrections, please email run@ebrun.com .

I am a Brand Seller Factory Service Provider Marketplace Seller Researcher Read it again.

2026年8月披露的信息显示,从商汤芯片业务分拆而出的AI芯片厂商曦望完成20亿元人民币融资,对应投后估值约200亿元人民币。此轮融资后,公司估值较上一轮接近翻倍。

曦望的拆分源自商汤2024年底推出的1+X战略,商汤此前已在芯片研发领域累计投入超11亿元。独立运营的曦望主打大模型推理芯片研发,联席CEO分别为前百度创始成员王湛、前AMD及昆仑芯芯片老将王勇,核心团队规模近200人,多来自头部科技及芯片企业。

曦望此前已量产两款芯片,S1云边视觉推理专用芯片累计出货超万片,NPU IP曾授权索尼小米等企业。S2大模型推理GPGPU于2024年量产,兼容CUDA架构,下一代S3芯片仍瞄准推理计算市场。

当前AI算力需求正从训练侧逐步向推理侧倾斜,2026年推理算力占全部AI计算能力的三分之二,整体市场空间突破3000亿元。国内科技企业正加大相关领域投入,打造大模型部署环节的国产替代方案。

本次融资获得的资金,将支撑曦望拓展产品研发及商业化运营相关布局。

本文首发于 亿邦动力 官方网站

文章来源:亿邦动力

广告
微信
朋友圈

FAQ回顾

曦望是一家什么企业?

曦望是商汤2024年底推出的1+X战略下从芯片业务分拆而出的AI芯片厂商,主打大模型推理芯片研发,联席CEO分别为王湛、王勇,核心团队规模近200人,成员多来自头部科技及芯片企业。

曦望已推出的AI芯片产品有哪些?

曦望已量产两款芯片,S1云边视觉推理专用芯片累计出货超万片,NPU IP曾授权索尼、小米等企业;S2大模型推理GPGPU于2024年量产,兼容CUDA架构,下一代S3芯片仍瞄准推理计算市场。

AI推理芯片市场的发展趋势是什么?

当前AI算力需求正从训练侧逐步向推理侧倾斜,2026年推理算力占全部AI计算能力的三分之二,整体市场空间突破3000亿元,国内科技企业正加大相关领域投入,打造大模型部署环节的国产替代方案。

曦望本次融资的资金用途是什么?

曦望本次完成20亿元人民币融资,投后估值约200亿元,较上一轮接近翻倍,本次融资获得的资金,将全部用于支撑曦望拓展产品研发及商业化运营相关布局。

这么好看,分享一下?

朋友圈 分享

APP内打开

赞 +1
+1
微信好友 朋友圈 新浪微博 QQ空间
关闭
收藏成功
发送
/140 0