“Is the continuous upgrade of AI liberation or slavery?”

Produced by | JiXin

Key points in October:

1. There were 90 investment and financing events in the field of AI this month, with a disclosed amount of 6.252 billion RMB.

2. There were 17 investment events with amounts exceeding 100 million RMB.

3. There were 4 IPO events in the field of AI in October.

4. Among all financing events, excluding equity investments, the highest proportion was angel round financing, with a total of 10 events, accounting for 11.1%.

01 There were 90 investment and financing events in the AI market in October, with a disclosed financing amount of 6.252 billion RMB

According to Xiniu Data, there were 90 investment and financing events in the field of AI in China in October 2024, an increase of 12 compared to September, but the disclosed financing amount decreased by 40%. A total of 50 financing amounts were not disclosed.

Excluding acquisitions and equity transfers, the largest event was the IPO of "Horizon Robotics", a provider of intelligent driving solutions for passenger vehicles, with an amount of HK$5.142 billion.

Compared to the same period last year, there were 3 more investment and financing events, with a slight decrease in financing amounts and a slight reduction in average financing amounts.

Figure 1: Number and disclosed amounts of investment and financing events in the field of AI in China from October 2023 to October 2024 (Data source: Xiniu Data, Tianyancha, etc.)

Among various sub-sectors, the number of investment and financing events in the field of AI application topped the list with 41 events, accounting for 45.6%; the number of financing events in the field of intelligent robots ranked second with 10 events, accounting for 11.1%; the field of autonomous driving had 9 events, ranking third with a share of 8.97%, an increase of 2 compared to September.

It can be seen that among the 90 events in October, 3 occurred in the field of embodied intelligence, a decrease of 2 compared to September. They are as follows:

A 300 million RMB Pre-A round of financing for "Xingdongjiayuan", a developer of embodied intelligence technology and general humanoid robots, led by Qingliu Capital, Yuanjing Capital, and Alibaba, with additional investments from Ceyuan Capital, Legend Capital, Century Jinyuan, Jinding Capital, Zeyu Capital, and Qingkong Tiancheng. A 10 million RMB angel round of financing for "Qingzhou Zhihang", a developer of embodied intelligence robots, invested by Tianronghaise, Wanwu Weichuangtou, and Kuake Electric Power. A tens of millions of RMB B round of financing for "Zhixing Robotics", a developer of dexterous hands and embodied intelligence robots, led by Chengmei Capital and Yarui Angel.

Another noteworthy event was the acquisition of a large AI content creation model:

"Waveform Intelligence", a developer of large AI content creation model technology, was acquired by OPPO.

Figure 2: Statistics on investment and financing events in various sub-sectors of the AI field in China in October 2024 (Data source: Xiniu Data, Qichacha, Tianyancha)

From the perspective of financing rounds, excluding equity financing, equity transfers, private placements on the New Third Board, and IPO events, the number of companies completing angel round financing was the highest among the 90 investment and financing events in October, with a total of 10 events, accounting for 11.1%, mainly in early-stage financing. There were 6 A/A+ round financing events, accounting for 6.7%, an increase of 2 compared to September.

The investment and financing performance in the mid-to-late part of October was good, starting from the B/B+ round, with 2 C/C+ rounds and 1 D/D+ round.

The B/B+ round, C/C+ round, and D/D+ round financing events are as follows:

B/B+ Round

A multi-hundred million RMB B round of financing for "Pingfanghe", a provider of industrial vision inspection systems, led by Cowin Capital, with follow-up investments from Yuanhe Puhua, Yahui Investment, Yunhui Capital, Xiangcheng Jinkong, Entropy One Capital, and Boxing Capital.

A tens of millions of RMB B round of financing for "Zhixing Robotics", a developer of dexterous hands and embodied intelligence robots, led by Chengmei Capital and Yarui Angel.

C/C+ Round

A 298 million USD C round of financing for "Didi Chuxing Autonomous Driving", a developer of autonomous driving technology, led by GAC Group (601238), with follow-up investment from Didi Chuxing (DIDI).

A multi-hundred million RMB C+ round of financing for "Lingsheng Technology", a provider of general autonomous driving solutions, invested by Zhulu Capital.

D/D+ Round

A 100 million RMB D+ round of financing for "Qianglianzhichuang", a provider of integrated solutions for intelligent diagnosis and treatment of cerebrovascular diseases, led by Shunxi Fund, with follow-up investment from Yizhuang Guotou.

Figure 3: Distribution of financing rounds for investment and financing events in the field of AI in China in October 2024 (Data source: Xiniu Data, Qichacha, Tianyancha)

There were 4 IPO events in the field of AI in October, an increase of 2 compared to September.

Figure 4: Statistics on IPO events in the field of AI in China in October 2024 (Data source: Xiniu Data, Qichacha, Tianyancha)

From the perspective of financing amounts, excluding undisclosed events, the number of multi-million RMB investment and financing events accounted for the highest proportion, with a total of 19 events, accounting for 21.1%. There were 16 events with financing amounts ranging from 100 million to 1 billion RMB and 1 event with an amount exceeding 1 billion RMB. Compared to September, the number of large-scale investment events increased significantly.

Left: Figure 5: Statistics on the proportion of financing amounts in the field of AI in China in October 2024 (Data source: Xiniu Data, Qichacha, Tianyancha)

Right: Figure 6: Statistics on investment and financing events with amounts exceeding 100 million RMB in the field of AI in China in October 2024 (Data source: Xiniu Data, Qichacha, Tianyancha)

02

The C-end market becomes the next direction of advancement, practicality is key | Review of key financing events in October

1. "Boson Quantum" completes hundreds of millions of RMB in Series A financing

This is the fifth round of financing that Boson Quantum has completed since its establishment over 3 years ago. This round of financing was led by well-known VC Qifu Capital, with follow-up investments from Amiba Capital, Yuanhe Capital, and Yingfutech. This financing will be mainly used for the following three aspects: Firstly, for the continuous research and development of coherent light quantum computers and the formation of small-batch production capacity; secondly, for the construction of the "quantum computing+" ecosystem and practical applications in various industries; and thirdly, for continuous investment in the research and development and layout of general-purpose optical quantum computer technology. (36Kr)

2. Qiniu Intelligence makes its debut on the stock market, opens lower by 41.82% in the morning session

On October 16, Qiniu Intelligence (02567) made its debut on the stock market. The announcement showed that the share price was set at HK$2.75 per share, with a total of 160 million shares issued, 1,000 shares per lot, and a net proceeds of approximately HK$367.2 million. As of press time, the share price had fallen by 41.82%, reporting HK$1.6, with a trading volume of HK$25.7732 million.

It is reported that Qiniu Intelligence is one of the earliest platforms in China to provide audio and video cloud services, committed to becoming a world-leading one-stop scenario-based intelligent audio and video service provider. The company mainly provides audio and video PaaS services (MPaaS) and application platform as a service (APaaS). It is one of the few companies in China that has developed comprehensive and high-performance MPaaS products and has the technical capability to integrate and cover all aspects of the audio and video business.

According to iResearch, in terms of revenue in 2023, the company was the third-largest audio and video PaaS service provider in China, with a market share of 5.8%; in terms of APaaS revenue in 2023, the company was also the second-largest audio and video APaaS service provider in China, with a market share of 14.1%. (ECNS)

3. "Xingdongjiayuan" completes nearly 300 million RMB in Pre-A round financing, focusing on the large model of native robots + AI hardware platform

On October 16, embodied intelligence company "Xingdongjiayuan" recently completed nearly 300 million RMB in Pre-A financing. This round was jointly led by Qingliu Capital, Yuanjing Capital, and Alibaba, with follow-up investment from Ceyuan Capital. Existing shareholders Legend Capital, Century Jinyuan, Jinding Capital, Zeyu Capital, and Qingkong Tiancheng continued their investments, and Huaxing Capital continued to serve as the exclusive financial advisor for this round.

The financing funds will be used to accelerate groundbreaking research and development in embodied intelligence technology and the commercialization process of general-purpose humanoid robots, and to continuously build the company's talent and technology barriers.

This is also the third round of financing that "Xingdongjiayuan" has completed since its establishment; previously, the company received over 100 million RMB in angel round financing led by Legend Capital, with follow-up investments from Jinding Capital, Zeyu Capital, Qingkong Tiancheng, and additional investment from existing shareholder Century Jinyuan, as well as tens of millions of RMB in seed round financing led by Century Jinyuan, with follow-up investment from Turing Venture Capital."Xingdongjiayuan" was founded by Assistant Professor Chen Jianyu from the Institute for Interdisciplinary Information Sciences at Tsinghua University. It is the only humanoid robot enterprise with shares held by Tsinghua University and is supported by the Shanghai QiZhi Institute. The company focuses on the research and development of native general-purpose embodied agents and has launched a new hardware platform combining native robot large models and AI definitions. (36Kr)

4. The field of embodied intelligence stirs up again, "Noematrix" completes hundreds of millions of RMB in Pre-A round financing, helping AI enter the real world

On September 6, Noematrix, a startup in the field of embodied intelligence, announced the completion of a cumulative Pre-A round of financing amounting to hundreds of millions of RMB.

Founded in November 2023, Noematrix was strategically incubated by Non-X Technology Group, which has been deeply involved in the field of general-purpose intelligent robots for nearly a decade. It is committed to the research and development of embodied intelligence systems and related tools and platforms.

This round of financing was jointly led by renowned investment funds Prosperity7 Ventures and GF Securities Capital, with participation from well-known investment institutions such as Zeyu Capital, Innovation Works, Qiji Chuangtan, Plug and Play China, and MFund.

Adhering to the business strategy of "technology-driven, product-first", Noematrix closely connects with the market demand for intelligent agents. In the early stages of its establishment, Noematrix officially launched its self-developed embodied brain, Noematrix Brain. Noematrix firmly believes that intelligent infrastructure with embodied brain systems as its core has huge commercial and social value in promoting technological innovation and product applications in the robotics industry. This system is designed with "universality" and "intelligence" as its core, providing a comprehensive, integrated software and hardware platform and tools covering the entire lifecycle of intelligent agent applications, from research and development, testing, to deployment and verification. (36Kr)

5. Pingfanghe completes hundreds of millions of RMB in Series B financing, led by Cowin Capital

On October 17, Pingfanghe (Beijing) Technology Co., Ltd. (hereinafter referred to as "Pingfanghe") announced the completion of hundreds of millions of RMB in Series B financing. This round of financing was led by Cowin Capital, with follow-up investments from Yuanhe Puhua, Yahui Investment, Yunhui Capital, Xiangcheng Jinkong Capital, Entropy One Capital, and Boxing Capital.

Pingfanghe is dedicated to intelligent solutions for computer vision systems, providing full-stack vision services for manufacturing enterprises, including optical imaging system design, custom visual algorithms, custom and development of visual equipment, and integration of visual and production data. It is the first company in China to propose the concept of industrial cloud vision. (Investing.com)

6. Deeply rooted in the application of AI large models in the telecommunications industry, "Wangsi Technology", a national specialized and new small giant enterprise, announces the completion of over 100 million RMB in Series A+ financing

National specialized and new small giant enterprise Wangsi Technology announced the completion of Series A+ financing exclusively invested by Fanshi Fund, with a cumulative financing amount exceeding 100 million RMB. The successful fundraising not only demonstrates Wangsi Technology's resilience and vitality in the capital market amidst adversity but also represents a firm endorsement by domestic investment institutions of Wangsi Technology's strength and future prospects. This round of funding will help Wangsi Technology further deepen its technological research and development, promote its expansion into more digital solution areas, and facilitate the integrated development of upstream and downstream sectors in the AI industry."With its forward-thinking digital mindset, high-quality solutions, and user experience at its core, Wangsi Technology has successfully assisted enterprises such as China Mobile, China Unicom, China Telecom, GAC Group, and Shandong Heavy Industry in their digital transformation, contributing to China's digital construction efforts. Its strong technical expertise and exceptional service have won it high praise and trust from customers. (Wangsi Technology)

7. "Gongji Technology" completes tens of millions of RMB in financing for developing a power grid-style computing power scheduling network

On October 21, "Gongji Technology" completed tens of millions of RMB in angel round financing, led by Shuimu Venture Capital, with additional investment from seed round investor Qiji Chuangtan and new venture capital. The financing was facilitated by FA services provided exclusively by Yuanhe. The funds will be primarily used for product research and development."It is understood that "Gongji Technology" is mainly committed to solving the problem of difficult access to computing power resources in elastic and temporary computing scenarios. It has built a power grid-style computing power scheduling network that utilizes low-cost idle computing power to serve high-value flexible demands."In terms of computing power supply, "Gongji Technology" has built the SLM task scheduling and distribution platform, achieving a global first in both idle time invocation and one-click access. The platform has integrated idle resources from a large number of personal devices, thousands of Internet cafes across the country, and medium to large clusters. (ECNS)

8. Qianglianzhichuang obtains over 100 million RMB in Series D+ financing, led by Beijing Pharmaceutical Industry Fund

On October 21, Qianglianzhichuang (Beijing) Technology Co., Ltd. (hereinafter referred to as "Qianglianzhichuang®") obtained over 100 million RMB in Series D+ financing, led by the Beijing Pharmaceutical and Health Industry Investment Fund (hereinafter referred to as "Beijing Pharmaceutical Industry Fund"), with follow-up investment from the Beijing Economic-Technological Development Area Industrial Upgrading Equity Investment Fund (hereinafter referred to as "Yizhuang Industrial Upgrading Fund").

8. Didi Autonomous Driving Completes $298 Million Series C Funding Round Led by GAC Group

On October 22, Didi Autonomous Driving announced the completion of its Series C funding round of $298 million. This round was led by GAC Group, with Didi participating. The funds will be used to increase investment in autonomous driving technology research and development, accelerating the mass production of the first Robotaxi.

Didi Autonomous Driving aims to utilize artificial intelligence technology to advance the research and product application of L4 autonomous driving technology. It has been operating safely in Beijing, Guangzhou, and Shanghai for over 1,500 consecutive days. In April this year, Didi Autonomous Driving and GAC Aion's joint venture, Andi Technology, received its business license. Both parties will deeply integrate leading autonomous driving technology with mature experience in new energy vehicle manufacturing, aiming to launch the first mass-produced L4 vehicle by 2025.

Zhang Bo, CTO of Didi and CEO of Didi Autonomous Driving, said that L4 autonomous driving technology can significantly reduce traffic accidents and improve traffic efficiency, making it one of the most important technological changes in the next decade of the AI era. The safe and stable development of autonomous driving requires breakthroughs and deep integration across AI, mass production, and operational networks. (The Beijing News)

9. Aptiv Announces Strategic Investment in MAXIEYE

On October 24, Aptiv, a global technology company committed to making mobility safer, greener, and more connected (NYSE: APTV), announced a final agreement to invest approximately RMB 570 million in MAXIEYE, a Chinese local company. After the investment, Aptiv will own approximately 18% of MAXIEYE, becoming its largest shareholder.

MAXIEYE is a Chinese local core technology service provider in intelligent driving and smart mobility, possessing self-developed BEV platform architecture and data intelligence systems. It provides full-condition, multi-scenario, cross-platform ADAS and ADS system products and solutions. Founded in 2016, MAXIEYE currently employs over 300 people, with over 70% being R&D personnel. MAXIEYE will become Aptiv's visual perception software technology partner in China, helping Aptiv further develop flexible, cost-effective, modular intelligent driving solutions and maintain a leading position in the rapidly growing Chinese ADAS market. (Aptiv China)

10. Horizon Robotics, the Largest Tech IPO in Hong Kong This Year, Lists on the Hong Kong Stock Exchange Raising HK$5.407 Billion

On October 24, Horizon Robotics, an intelligent driving technology enterprise, listed on the main board of the Hong Kong Stock Exchange, raising a total of HK$5.407 billion (IT Home Note: approximately RMB 4.961 billion), making it the largest tech IPO in Hong Kong this year.

On its first trading day, Horizon Robotics-W opened at HK$5.12, surging 28.32%.

The prospectus revealed that Horizon Robotics recorded revenues of approximately RMB 467 million, RMB 906 million, and RMB 1.552 billion in 2021, 2022, and 2023, respectively, with a compound annual growth rate of 82.3%. In the first half of 2024, revenues reached approximately RMB 935 million, a year-on-year increase of 151.6%.

According to official information, Horizon Robotics' hardware-software integrated solutions have been adopted by 27 OEMs (42 OEM brands) and are used in 290 vehicle models. All of China's top ten OEMs have chosen Horizon Robotics' intelligent driving solutions. (IT Home)

11. Receiving Tens of Millions in Series B Funding, "ZhiXing Robotics" Accelerates Its Layout in Embodied Intelligence

On October 29, ZhiXing Robotics Technology (Suzhou) Co., Ltd. (hereinafter referred to as "ZhiXing Robotics") completed tens of millions of yuan in Series B funding. This round was jointly led by Chengmei Capital and Zhongguancun Zhiyou Scientists Fund. The funds will be used for the research, development, expansion, and promotion of the company's core products and systems.

Founded in August 2018, ZhiXing Robotics has rapidly developed into a national high-tech enterprise focused on the research and production of dexterous hands and embodied intelligent robots.

The company is dedicated to developing flexible, reliable, safe, and intelligent robotic hands and robotic system solutions. To date, it has served over 300 customers and established cooperation with over 50 domestic enterprises, accumulating successful cases in intelligent manufacturing and automation across multiple industries. (36Kr)

12. Focusing on "AI + Tourism", "Vision Travel Technology" Receives Hundreds of Millions of Yuan in Series A Funding

On October 30, "Vision Travel Technology" announced the completion of hundreds of millions of yuan in Series A funding. Investors include renowned institutions such as Shenbao Yiben Fund, Guangzhou City Investment Group's Guangdong-Hong Kong-Macao Greater Bay Area Cultural Industry Investment Fund, Gaorong Venture Capital, Youshan Investment, and Zhonghu Yakang, an investment platform of Yakang Stock. The total funding exceeds RMB 100 million, with Huaxing Capital serving as the exclusive financial advisor for this round. The funds will be used to accelerate the research and development upgrade of the VtripGPT tourism big data model and continuously promote its market application in high-quality domestic and international tourist destinations.

Founded in November 2021, "Vision Travel Technology" (hereinafter referred to as "Vision Travel") aims to build a "new generation of AI-driven tourism platforms" based on big data models. It addresses various pain points faced by consumers during travel, such as cumbersome itinerary planning, redundant pricing and strategy information, popular itineraries, lack of in-depth experiences, and insufficient high-quality sharing materials. Currently, its C-end user-oriented products such as "Lepai Network" and "Golden Panda Travel" can achieve functions like AI itinerary design, dynamic package booking, and intelligent travel photography. (36Kr)

03 Corporate Strategic Layout Drives Equity Transactions, Embracing Innovation Through Acquisitions | Monthly Mergers and Acquisitions Report

1. Sanfeng Intelligence Plans to Acquire Shares in Its Subsidiary Through Its Wholly-Owned Subsidiary Xinyanlong

On October 11, Sanfeng Intelligence announced that its wholly-owned subsidiary, Shanghai Xinyanlong Automobile Equipment Manufacturing Co., Ltd. (hereinafter referred to as "Xinyanlong"), plans to use its own funds to acquire a 45% stake in its controlled subsidiary, Sanfeng Zhilian, held by Mr. Wang Weixin, for a total equity transfer price of RMB 3 million.

After the completion of this equity acquisition, Xinyanlong will hold a 45% stake in Sanfeng Zhilian, increasing Sanfeng Intelligence's total holding in Sanfeng Zhilian from 55% to 100% through direct shareholding and through Xinyanlong.

Public information shows that Sanfeng Intelligence is a supplier and integrated service provider of intelligent manufacturing system solutions and equipment with industrial robots and related intelligent technology as its core. Its main business includes five major segments: intelligent conveying equipment, intelligent welding production lines, industrial mobile robots and intelligent three-dimensional warehousing system equipment, high and low voltage equipment and electrical control equipment, intelligent parking equipment, and other businesses such as oil product trading.

Sanfeng Intelligence stated that future production process equipment will shift towards intelligent flexibility and AI unmanned manufacturing, and the demand for high-end intelligent equipment in automobile manufacturing is expected to maintain a growth trend, which will also drive the steady growth of the company's automobile intelligent equipment business scale. The company will continuously strive to improve its profitability through technological and product innovation and upgrades. (Gaogong Robot)

2. Giant Companies Make Substantial Premium Acquisitions, and Sungrow Power Supplies Intends to Acquire Taihe Intelligence

On October 18, Sungrow Power Supplies' wholly-owned subsidiary, Sungrow Renewable Energy Development Co., Ltd. (hereinafter referred to as "Sungrow Renewable Energy"), signed a share transfer agreement with Xu Dahong, Ge Suhui, Tang Lin, and Wang Jincheng regarding Hefei Taihe Intelligent Technology Group Co., Ltd. On the same day, Sungrow Renewable Energy also signed a voting rights waiver agreement with Xu Dahong and his concerted action person Yang Yalin, as well as voting rights entrustment agreements with Ge Suhui and Wang Jincheng.

It is reported that the target company is a provider of complete sets of intelligent equipment and services based on AI visual recognition. After years of development, it has accumulated certain expertise in spectral detection technology, intelligent algorithm technology, and industrial robot automation technology. In recent years, it has actively expanded into the new energy business field using its technological accumulation. Based on business needs, the company intends to obtain control of the target company through this transaction, giving full play to the target company's advantages in spectral detection technology, intelligent algorithm technology, and industrial robot automation technology, thereby helping the company further improve production efficiency, product yield, and automation and intelligence levels. (ECNS)

3. OPPO Acquires Wave Intelligence, an AI Company Founded by Post-90s Entrepreneurs, Which Has Already Received Tens of Millions in Funding

According to multiple sources on October 22, Hangzhou-based AI startup Wave Intelligence will be acquired by smartphone manufacturer OPPO, with acquisition negotiations recently completed.

Multiple sources confirmed that the news is basically true, and Wave Intelligence's founder and CEO Jiang Yuchen, along with the company's core members, will join OPPO. However, the company has not yet completed the employee agreement settlement and is still in the negotiation stage. It should be noted that the news about "Wave Intelligence dissolving" is untrue.

In terms of products, during the Pre-A funding round in January 2024, Wave Intelligence released its self-developed large Chinese creative domain model "Weaver," jointly developed by Wave Intelligence, APUS, and Professor Chen Huajun's team at Zhejiang University. It also launched an AI-driven user writing agent product, "Wawa Writing 1.0," making Weaver the first domestic infinite long-text generation large model. In July this year, Wave Intelligence released a new generation of personalized adaptive private language model for multimodal infinite long-content generation, "Weaver 2.0," as well as a series of products such as the multilingual and multimodal AI content creation tools "Wawa Writing 2.0," "Siuuu.AI," and "AI Learning." (TMTPost)

04 Significant Developments on Mobile Platforms, with More Specialized AI Capabilities | Review of October's Key Tech News

Swipe up to read event details

1. OpenAI Launches New Writing and Coding Tool ChatGPT Canvas, Which Will Be Freely Accessible

On October 4, OpenAI announced the launch of a new writing and coding tool, Canvas, a new interface for collaboratively writing and coding projects with ChatGPT, going beyond simple chatting.

2. MediaTek's New Generation Dimensity Flagship Chip Is Optimized for Google's Gemini Nano Multimodal AI, Supporting Image and Audio Processing

On October 8, MediaTek announced that its new generation Dimensity flagship chip has been optimized for Google's large language model, Gemini Nano. The company stated that the upcoming Dimensity flagship chip is optimized for Google's Gemini Nano multimodal AI, supporting not only text but also image and audio processing.

3. vivo Launches New BlueHeart Large Model Matrix, Introducing the 3B and Voice Large Models with 3 Billion Parameters

On October 10, during the opening speech of the 2024 vivo Developer Conference, vivo officially launched its self-developed BlueHeart large model matrix, comprehensively upgrading language and end-side large model capabilities. It also introduced vivo's self-developed voice, image, and multimodal large models.

4. ByteDance Releases GR-2 Robot AI Large Model: Average Task Completion Rate of 97.7%, Simulating Human Learning to Handle Complex Tasks

On October 10, ByteDance's research team launched the second-generation robot large model, GR-2 (Generative Robot 2.0). Its highlight is the innovative construction of a "robotic infancy" learning phase, mimicking human growth and learning complex tasks, possessing excellent generalization ability and multitask versatility.

5. Major Announcement! MileLaw Officially Launches the Worry-Free AI Legal Service Platform

On October 10, MileLaw officially launched the Worry-Free AI Legal Service Platform, which is based on MileLaw's self-developed legal large model with 10 billion parameters. This enterprise AI legal service platform provides intelligent Q&A, contract review, document drafting, and other AI legal services.

6. Microsoft Drives AI Transformation in the Healthcare Industry: Introduces New Agent Services and Expands Azure Large Models

On October 11, tech media NeoWin reported that Microsoft had launched a new medical agent service for Copilot Studio to address numerous challenges in the current healthcare system.

7. Tesla Unveils New Products at Its Event, Including Vehicles Without Steering Wheels or Pedals

On October 11, Tesla held a "Robotaxi Day" event in California with the theme "We, Robot." At the event, new products such as the autonomous taxi Cybercab, Robovan (autonomous van), and Optimus robot were unveiled.

8. The First AI Kaggle Grandmaster Is Born: OpenAI's o1-preview Wins Seven Gold Medals

On October 12, tech media The Decoder reported that OpenAI had launched the MLE-bench, a new benchmark designed to evaluate the ability of AI agents in developing machine learning solutions. This benchmark includes 75 Kaggle competitions, aiming to measure the progress of autonomous AI systems in machine learning engineering. These competitions cover various fields, including natural language processing, computer vision, and signal processing.

9. Kuaishou Jointly Launches the Pyramid Flow Open-Source Text-to-Video AI Model: Up to 10 Seconds at 768P@24fps, Innovative Pyramid Flow Matching Reduces Inference Tokens by a Quarter

On October 12, a research team jointly formed by Kuaishou Technology, Peking University, and Beijing University of Posts and Telecommunications launched the Pyramid Flow text-to-video model. After inputting text, users can generate high-quality videos up to 10 seconds long, with a resolution of up to 1280x768 and a frame rate of 24fps.

10. China Mobile Launches the Jiutian Shanzhi Multimodal Base Large Model and Over 30 Self-Developed Industry Large Models

On October 12, IT Home learned from the China Mobile official website that at the main forum of the 12th China Mobile Global Partner Conference, China Mobile launched the Jiutian Shanzhi multimodal base large model and introduced over 30 self-developed industry large models.

11. CogView3-Plus, an open-source text-to-image model from Zhipu, launches related features on the Zhipu Qingyan App

On October 14, the Zhipu technical team announced the open-source launch of the CogView3 and CogView3-Plus-3B text-to-image models. The capabilities of these models are now available on the "Zhipu Qingyan" App.

12. Microsoft explores AI-driven audio-to-image models for real-time visualization of conference speakers

On October 15, tech media MSPoweruser reported that Microsoft has obtained a new patent describing the generation of images based on real-time user voice input. According to the latest listing by the U.S. Patent and Trademark Office, the 20-page patent was filed by Microsoft on April 5, 2023, and approved on October 10. The patent describes a system that can capture audio in real-time during meetings or lectures, summarize it through a language model, and generate corresponding AI images.

13. Alibaba Cloud International releases Marco, a large-scale translation model supporting 15 mainstream languages

On October 16, Alibaba Cloud International's AI team AIDC-AI announced the official launch of Marco, a newly developed large-scale translation model. Marco supports 15 global mainstream languages, including Chinese, English, Japanese, Korean, Spanish, and French. Based on the widely used open-source evaluation dataset Flores, Marco outperforms benchmark translation products like Google Translate, DeepL, and GPT-4 in BLEU automatic evaluation metrics.

14. Adobe introduces multiple AI tools for creating 3D scenes, removing bystanders, and cleaning up footage

On October 16, The Verge reported that Adobe recently showcased several experimental AI tools for animation, image generation, and photo and video optimization, which are expected to be integrated into Creative Cloud in the future.

15. Zuckerberg: Lenovo builds personal AI agents AI Now based on Meta's Llama large model

On October 15 (U.S. Western Time), Lenovo held its annual Tech World conference in Seattle, USA. During the keynote speech, Lenovo CEO Yang Yuanqing and Meta founder and CEO Mark Zuckerberg announced the launch of AI Now, a personal AI agent for PCs based on Meta's Llama large model, in collaboration with Lenovo.

16. ZeroOneThing releases its new flagship model Yi-Lightning: LMSYS tests show it surpasses GPT-4 in China, ranking first nationally and sixth globally

On October 17, ZeroOneThing, an AI company founded by Kai-Fu Lee, released its new flagship model Yi-Lightning. On the blind test list LMSYS, Yi-Lightning surpassed GPT-4o-2024-05-13 and Claude 3.5 Sonnet released by OpenAI and Anthropic, ranking sixth globally and first in China.

17. Rise of edge AI: Mistral releases Ministral 3B and 8B, claimed to be 'the best edge models in the world'

On October 17, Mistral introduced two new AI models, Ministral 3B and Ministral 8B, designed to work without the need for cloud server connections, enhancing local AI experiences on devices like laptops and smartphones. The company claims they are 'the best edge models in the world'.

18. Microsoft open-sources bitnet.cpp, a 1-bit LLM inference framework that can run AI models with 100 billion parameters locally without a GPU, reducing energy consumption by up to 82.2%

On October 19, tech media marktechpost reported that Microsoft had open-sourced bitnet.cpp, a highly efficient 1-bit Large Language Model (LLM) inference framework that can run directly on CPUs. Users can run LLMs with 100 billion parameters on local devices using bitnet.cpp without a GPU, achieving a 6.17x speed-up and reducing energy consumption by up to 82.2%.

19. China's first large clinical ophthalmology model, VisionFM, launched for detecting and predicting eye health status

On October 21, IT Home learned from Henan News that VisionFM, China's first large clinical ophthalmology model, was launched in Zhengzhou on October 17. According to Wang Ningli, President of the Henan Academy of Medical Sciences, VisionFM is both a cutting-edge platform and an AI application. It boasts a powerful "brain" comprising 3.4 million eye images from 560,000 individuals in 26 countries and regions, making it the world's largest ophthalmology image database.

20. 'Pure-blooded HarmonyOS' ecosystem improves with smarter AI assistants; public beta testing for more devices to commence

On October 22, at the HarmonyOS Night and Huawei Full Scenario Launch Event, Huawei announced the official release of its first domestic mobile operating system, HarmonyOS. The new OS integrates 'AI+OS' for more efficient AI model operation. Huawei also unveiled an upgraded AI assistant named "Xiaoyi," which integrates native AI capabilities into the OS for the first time. Powered by the Pangu large model, "Xiaoyi" has enhanced perception and reasoning abilities, supports 23 top-tier scenario memory perceptions, and achieves a task success rate exceeding 90% with over one trillion pieces of knowledge.

21. IBM launches Granite 3.0: The ultimate enterprise AI model, laying a solid foundation for AI agents

On October 22, tech media NeoWin reported that IBM had launched the new Granite 3.0 series of AI models at its annual TechXchange event. These models can match or surpass models of similar scale in academic and industry benchmarks.

22. Honor MagicOS 9.0 upgrades to support a 3 billion-parameter edge large language model, reducing power consumption by 80% and memory usage by 1.6GB

On October 23, Honor officially launched MagicOS 9.0, billed as the "industry's first personal AI operating system with agents for all scenarios."

23. JetBrains creates Mellum, the ultimate AI assistant for developers, offering low latency, fast completion, and high accuracy

On October 23, JetBrains announced the launch of Mellum, a new large language model designed to provide faster, smarter, and more context-aware code completion for software developers. According to the company, Mellum stands out from other large language models by being specifically designed for developers, offering low latency, strong performance, and comprehensive functionality, providing relevant suggestions in minimal time.

24. OPPO unveils its annual flagship Find X8 series, setting a new benchmark for imaging and AI capabilities

On October 24, OPPO released its new annual flagship Find X8 series. In the field of AI, the Find X8 series introduces a new one-click screen inquiry feature, allowing users to easily activate it with a single step and perform object recognition, Q&A, and navigation based on screen content. It also uses AI to remove reflections, motion blur, and unwanted objects, enhancing photo quality. Additionally, in the office sector, AI can generate copywriting and extract documents, enhancing users' productivity.

Huawei, Xiaomi, Honor, OPPO