2026-10-18 16:02

Introduction: The Momenta chip team is expanding into external markets driven by large model demand. Starting from automotive ADAS, it now targets broader AI computing needs—this positioning of XINXIN HANGTU offers a lens to observe the current wave of semiconductor entrepreneurship: teams with expertise in algorithms, products, and system R&D are actively transforming their accumulated knowledge into chip products, broadening customer base and application scope.

Starting from automotive ADAS, targeting broader AI computing demands, XINXIN HANGTU’s strategic positioning provides a critical entry point for observing this new wave of chip startups: teams with deep algorithmic, product, and system development experience are now translating their existing expertise into silicon solutions, expanding both client reach and application domains.

A fresh round of funding has brought XINXIN HANGTU, a physical AI chip company, and its star investor lineup into the spotlight.

Market reports indicate that XINXIN HANGTU recently closed a strategic financing round, with Renault Group making its first investment, Chery Automobile increasing its stake, and Longxin Technology joining as a new backer. The Star Daily on SciTech reporters noted that within less than three years since founding, XINXIN HANGTU has now raised approximately 5 billion RMB in cumulative funding.

In fact, besides XINXIN HANGTU, other emerging chip players such as Xiwang, Qianhe Yibang, and Lanxin Computing have also secured substantial funding in primary markets.

It’s evident that new demands driven by AI R&D and applications are opening doors for teams with industrial experience to independently develop chips and enter external markets.

Star Investor Consortium

Before this latest funding round, XINXIN HANGTU already had an extensive shareholder roster.

Business registration data shows that to date, XINXIN HANGTU has attracted dozens of external institutional investors, including multiple national-level funds, local state-owned capital platforms, automotive industry capital, and leading market-oriented institutions.

Among them, state-backed funds include the National Artificial Intelligence Industry Investment Fund (entered in early January), two science and technology equity funds funded by the Social Security Fund—Social Security Fund Yangtze River Delta Science & Technology Innovation Equity Investment Fund and Social Security Fund Bay Area Science & Technology Innovation Equity Investment Fund—alongside Guokai KeChuang, Guangzhou Industry Control Industrial Investment, Shenzhen Capital Venture, and Suzhou Capital Venture.

Industry capital includes Shangqi Capital (affiliated with SAIC Motor), Junsheng Electronics, NIO Capital, and Chery Automobile. The newly joined Renault Group and Longxin Technology are not yet listed in official business registry shareholder records. Market-oriented institutions encompass IDG Capital, Shunwei Capital, GGV Capital, CDV Investments, Zhenge Fund, and Cathay Capital.

The entry of industrial players like Longxin Technology undoubtedly elevates market attention toward this chip firm established under three years.

From its founding team perspective, XINXIN HANGTU maintains close ties with Momenta, a smart driving company recently listed on HKEX. Li Jun, legal representative and chairman of XINXIN HANGTU, was formerly head of Momenta’s chip division, with long-term experience in autonomous driving chip R&D. Additionally, key members of Momenta’s founding team—Cao Xudong and Xia Yan—are indirectly holding shares through XINXIN HANGTU’s controlling shareholder.

Both parties also collaborate commercially. In August this year, Momenta, XINXIN HANGTU, and QNX (a subsidiary of BlackBerry) jointly announced the establishment of an autonomous driving platform, integrating Momenta’s ADAS solution, XINXIN HANGTU’s X7 chip, and QNX’s secure operating system to deliver mass-production-ready solutions for automakers.

The aforementioned X7 chip is currently XINXIN HANGTU’s flagship product. Public information indicates that the X7 is an intelligent driving chip targeting high-end urban ADAS and L3/L4 autonomous driving markets. Currently, the X7 is already integrated into certain versions of SAIC Volkswagen ID.ERA 9X and MG07, and the upcoming Roewe Jia Yue 07, which has entered pre-sale, has confirmed adoption of this chip.

Notably, Longxin Technology, one of this round’s investors, has already mass-produced and commercialized its automotive-grade LPDDR5X memory products.

As large models are increasingly deployed at the edge, bandwidth, power consumption, overall solution cost, and supply chain stability have become shared concerns for ADAS mass production. This potential collaboration between the two companies is a major focus of market interest.

XINXIN HANGTU stated that bringing in Longxin Technology could provide advantages in memory pricing and supply. Another industry insider told The Star Daily on SciTech that, based on their respective business strategies, this equity partnership may enable Longxin Technology to further penetrate the automotive memory market, participate in computing platform integration, and engage in customer validation.

Previously, another industrial shareholder, Junsheng Electronics, had already initiated joint development with XINXIN HANGTU. According to Junsheng Electronics’ 2025 annual summary, the company developed a high-end intelligent ADAS domain controller based on XINXIN HANGTU’s chip platform; related chips have been successfully powered up and tested on vehicles.

New Chip Opportunities Driven by AI Applications

XINXIN HANGTU’s ambitions may extend beyond ADAS chips. In public presentations, the company positions itself as delivering hardware infrastructure for physical AI applications—such as autonomous driving and robotics—through dedicated chips designed specifically for large model workloads.

Starting from automotive ADAS and expanding toward broader AI computing needs, XINXIN HANGTU’s positioning offers a window into this new wave of semiconductor entrepreneurship: teams with proven expertise in algorithms, product design, and system development are now converting their accumulated capabilities into tangible chip products, aiming to serve wider markets and diverse applications.

Several companies that recently secured large-scale funding share this industrial background. Formerly embedded in OEMs, automotive chip divisions are now spinning off to raise external capital and pursue external supply expansion. Meanwhile, more R&D teams from AI and internet giants are entering the market: Xiwang, a reasoning GPU startup, originated from SenseTime’s chip division and was spun off independently at the end of 2024; Qianhe Yibang, a 3D memory-compute co-integrated chip company, was incubated by NetEase, with its core team having decades of experience in 3D integrated computing architecture; Lu Shan, founder of Lanxin Computing, a RISC-V server CPU company, previously led server chip R&D at ByteDance.

These teams possess deep technical accumulations closely aligned with real-world computing demands from application enterprises. As large models enter game development, enterprise services, and intelligent edge devices, increased frequency and scale of model inference bring rising operational costs—chips must now accommodate not only response speed and energy efficiency but also memory access patterns and system scheduling requirements.

For example, Xiwang focuses on inference GPUs to reduce post-deployment computational costs; Qianhe Yibang explores 3D DRAM stacking technologies to enhance data access efficiency between compute and memory; Lanxin Computing integrates AI acceleration units directly into server CPUs to address hybrid workloads combining general-purpose computing and AI tasks.

In April, Xiwang announced a funding round exceeding 1 billion RMB, with participation from Hangzhou Capital and others; in August, Qianhe Yibang disclosed a B-round funding exceeding 2 billion RMB, backed by China State Development Fund, China Mobile Chain Leader Fund, and Shanxing Capital; in October, Lanxin Computing revealed a series of B-round financing totaling nearly 2 billion RMB, with continued follow-on investments from existing backers like Shenzhen Capital Venture, and new participation from Suzhou Capital Venture and GD Microelectronics.

Industrial synergies have become crucial resources for market expansion among these firms. Xiwang has partnered with Youzu Games on AI content generation and real-time inference in game development; NetEase Youdao and China Mobile Chain Leader Fund provide business use cases and validation scenarios for Qianhe Yibang; as an early-stage investor’s industrial partner, Lenovo has launched servers equipped with Lanxin Computing’s chips and established a joint lab with the company.

A semiconductor industry expert told The Star Daily on SciTech: “Application-driven chip design is not new. What’s different this time is that model capabilities, computational workloads, and deployment paradigms are evolving rapidly, leaving room for new optimizations in chip architecture and hardware-software co-design. This enables teams originating from application companies to leverage their deep understanding of specific needs and combine it with existing R&D strengths to build products serving broader customer bases.”

Source: The Star Daily on SciTech ©