news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Today's Financing Express

Universal Robots Company Zhuji Power Completes RMB100 Million Series A Financing:

Zhuji Power is a general robotics company that focuses on the research and development and manufacturing of embodied intelligence and legged robots. Its products include humanoid bipedal and quad-legged robots and related software and hardware solutions. Founder Zhang Wei is a tenured professor at the Southern University of Science and Technology and was previously a tenured professor at the Department of Electrical and Computer Engineering at Ohio State University. This round of financing was led by Alibaba, China Merchants Venture Capital, and Shangqi Capital under SAIC Group, with original shareholders Frees Capital, Oasis Capital, and Vision Capital following suit. This is also Alibaba's first humanoid robot company.

AI data reliability startup Soda Data raises $14 million:

Soda Data's flagship product is a set of data quality tools called "SodaGPT". It combines the natural language processing skills of generative AI with the domain-specific language capabilities of Soda Checks Language or SodaCL to convert users' queries into simple data quality tests. This round of financing was led by existing investors and strategic partners Singular and Point Nine.

Industrial AI product developer Jingyun Intelligence completed tens of millions of yuan in Pre-A round of financing:

Founded in 2022, Jingyun Intelligence is an industrial brain and manufacturing digital transformation service provider, focusing on industrial Internet PaaS platform technology research and industrial AI product development. This round of financing was led by Yuanhe Yuandian and followed by Qianrong Venture Capital. The funds raised will be used to continuously optimize the core product "Jingyun OS Industrial Brain" and focus on the research and development of industrial AI products for segmented process scenario applications.

Woye Technology, a smart tour guide robot for scenic spots, received tens of millions of RMB in strategic investment:

Woye Technology is a developer of intelligent tour guide robots for scenic spots, focusing on the research and development, production and sales of intelligent robots. The company has mature core technologies such as multi-sensor identification and positioning autonomous navigation technology, deep neural network artificial intelligence technology, and robot sensor drive modularization. Recently, Woye Technology received a strategic investment of tens of millions of RMB from Jiangsu Guangmang Group.

Construction robot Fontan Robotics received tens of millions of strategic investment:

Fengtan Robotics is a construction robot company and a professional intelligent construction product and solution provider. Its main product is an indoor spraying robot with putty and latex paint functions. It recently received a strategic investment of tens of millions from Century Golden Resources.

Robotics developer Elythor receives strategic investment:

Swiss robotics developer Elythor has received strategic investment from InnoBooster. Founded on January 1, 2023, Elythor is a company focused on using multi-rotor aircraft, fixed-wing drones and flight deformation technology to develop agile robots capable of inspecting power infrastructure in complex spaces and over long distances.

Generative AI emoticon package Easel AI received $2.65 million in financing:

Easel AI is an AI image generation APP that is tied to IOS iMessage. It was founded in December 2023 in LA, USA. It is positioned as an AI image generation startup with a total of 7 employees and is in its infancy. One of the co-founders and CEO, Rajan Vaish, graduated from Stanford University with a postdoctoral degree and served as chief research scientist at the Snap Research Center. Another founder, Sven Kratz, also served as a principal researcher at Snap, promoting AR technology research, APP development and maintenance, etc. The two resigned from Snap at the same time and founded Easel AI together, serving as CTO and responsible for technology development.

Easel has raised a total of $2.65 million in seed funding from Corazon Capital, Unusual Ventures and several angel investors. The company was selected for the first Google Startup Accelerator program.

(Welcome to add WeChatAIyanxishe2, learn more about AIGC and financing, and chat with like-minded friends about new AI products)

Today's big factory news

Google Eureka AI model exposed in advance, Gemini is about to release five new features

Google is about to launch a new AI model called "Eureka", which performs well in natural language generation. The model has already appeared on LMSYS Arena. It is expected to be officially released on July 18. In addition to Eureka, Google Gemini is about to release new features, including Imagen3, Gemini customized GPT, etc., and will also launch personalized responses, scheduled reminders, recording and Google Photos integration. Google is stepping up the recruitment of beta testers for the iOS version of Gemini, and the iOS update may be released soon.

Baidu launches Wen Xiaoyan AI digital human social APP

Baidu recently launched an AI digital human social APP called "Wen Xiaoyan". Using the Wenxin model, users can communicate and interact with AI virtual characters in real time, establish emotional connections, and enhance the interactive experience. Each AI digital person provides a unique chat service and can become the user's encyclopedia, life assistant, and even spiritual mentor. When replying, the digital person displays voice and text, and enhances the sense of reality through body language.

Google releases Magic Insert: One-click drag and drop allows people to perfectly blend into new backgrounds

Magic Insert first fine-tunes the pre-trained text-to-image diffusion model using LoRA and learned text tags and fuses it with the CLIP representation of the target style. Using the Bootstrapped Domain Adaptation technique, the photorealistic object insertion model for a specific domain is adapted to a diverse domain of artistic styles. This method allows choosing between the degree of stylization and the fidelity to the details of the original subject, and can even introduce more novelty in the generation.

Try it out: https://magicinsert.github.io/demo.html

Project: https://magicinsert.github.io

Baidu Smart Cloud and Shandong University of Finance and Economics develop a new liberal arts education and teaching proprietary large model "AI Talent"

The "AI Talent" education and teaching model was jointly developed by Shandong University of Finance and Economics, Baidu Smart Cloud, and Shandong Yinrui Technology. The model can build an intelligent science and education environment through key technologies such as big data, machine learning, and knowledge graphs. It can also structure professional teaching materials, scientific research data, internship training cases, and other knowledge on the model platform to form its own knowledge base.

todayProduct News

Product HuntHot List，Simplify the creative process with AI Web Designer

AI Web Designer simplifies the web design process through AI. Users only need to enter the website domain name to get a new design generated by AI. The platform not only provides free homepage design generation services, but also supports design editing, Figma export, and raw HTML code acquisition. It allows users to choose to obtain the design of the entire website, while providing white label services for agencies, enabling them to sell the design as their own products.

https://aiwebdesigner.io/?ref=producthunt

GitHub TrendingHot List，AI Research Report Creation Storm

The Storm system, which was open-sourced by the Stanford University research team, uses large language model technology to automatically research specific topics and generate complete reports with citations. Storm creates high-quality articles similar to Wikipedia through two stages: pre-writing and writing, combined with Internet search and multi-angle questioning. Storm's flexibility is reflected in its support for multiple search engine APIs and language models, which users can customize according to their needs. The system not only provides powerful tools for researchers and content creators, but also opens up new possibilities in the field of AI-assisted writing.

https://storm-project.stanford.edu/

Developer Recommendation

1. ReadLecture turns video content into essential notes instantly

ReadLecture is an AI tool for converting videos to text, which can greatly improve the efficiency of users watching videos. It only takes 5 minutes to understand the full content of a 2-hour lecture video. This tool incorporates cutting-edge technologies such as large language models, speech-to-text, and machine vision. It can accurately transcribe speeches and extract key frames (such as lecture PPTs), and provide detailed summaries, translations, and thinking notes.

The tool provides three output formats, suitable for immersive reading, quick browsing, and in-depth analysis. ReadLecture can also provide content summaries in various forms, such as mind mapping, question-and-answer understanding, and dialectical thinking.

www.readlecture.cn

2. SmartCrawl: Use AI technology to turn any website data into an API

Mendable.ai and the Firecrawl community have teamed up to launch the open source tool mendableai/firecrawl, which provides AI developers with powerful website content processing capabilities. The tool can convert the entire website into Markdown format or structured data suitable for large language models, and implement web crawling, crawling, and data extraction functions through a single API. Although it is still in the early stages of development, it has provided multiple SDKs and integration support with mainstream AI tools. Developers can use Python or Node SDK to easily implement functions such as website crawling, data extraction, and content search.

https://github.com/mendableai/firecrawl

3. PDF-Extract-Kit: AI-driven new PDF content extraction tool

The Opendatalab team launched PDF-Extract-Kit, an open source toolkit designed to improve the quality and efficiency of content extraction from PDF documents. This comprehensive toolkit uses LayoutLMv3 for layout analysis, YOLOv8 for formula detection, UniMERNet for formula recognition, and PaddleOCR for text recognition. PDF-Extract-Kit was designed with a variety of document types in mind, from academic papers to financial statements. It performs well when dealing with documents with poor scan quality or watermarks, showing strong adaptability.

https://github.com/opendatalab/PDF-Extract-Kit

4. AI video conversion: Live2Diff realizes real-time streaming processing

A research team from Shanghai Artificial Intelligence Laboratory, Saarland Computer Science Center, MaxPulter Institute of Computer Science, and S-Lab of Nanyang Technological University jointly developed an innovative technology called Live2Diff, which brings a major breakthrough in the field of real-time video stream conversion. This technology cleverly applies the unidirectional attention mechanism to the video diffusion model, achieving high-quality real-time video processing without relying on future frames.

Live2Diff can achieve a processing speed of 16FPS on the RTX 4090 GPU, demonstrating excellent performance. By adopting an efficient denoising scheme and pipeline processing, the technology surpasses existing methods in terms of temporal smoothness, efficiency, and generation quality.

https://live2diff.github.io/

Special attention

55% of OpenAI's annual revenue of $3.4 billion comes from C-end paying users

Research organization FutureSearch released a report on OpenAI's revenue. The report shows that OpenAI's annual recurring revenue reached 3.4 billion US dollars, of which more than half came from ChatGPT membership fees, while API revenue from enterprises and developers accounted for only about 15%.

ChatGPT Plus paying users account for 55% of OpenAI's revenue, or $1.9 billion, from 7.7 million subscribers worldwide, each paying $20 per month.

ChatGPT Enterprise accounted for 21% of total revenue, or $714 million, from 1.2 million enterprise users, each paying $50 per month.

API revenue accounted for 15%, or $510 million, from developers and companies using the ChatGPT API.

ChatGPT Team accounted for 8% of total revenue, or $290 million, from 980,000 subscribers, each paying $25 per month.

https://futuresearch.ai/request-openai-report

Stay tuned for the latest updates tomorrow!

AI Intelligence Agency is looking for intelligence partners to collect exclusive valuable clues! If you can provide information about the latest AI achievements, industry insider information, and unique products, please add the operation WeChat account:AIyanxishe2Note the industry position.

StandardBots received $63 million in financing; 7 key people behind Kuaishou Keling were exposed; OpenAI's top-secret project was exposed, and the nondisclosure agreement was reported again丨AI Intelligence Bureau

Groq receives $300 million investment from BlackRock, with latest valuation of $2.2 billion; Hayden AI receives $90 million in Series C funding丨AI Intelligence Agency

Skild AI raises $300 million to develop robot brains; Keling AI has 700,000 users; Alitong Yiqianwen ranks first in China, surpassing a number of open-source and closed-source models丨AI Intelligence Bureau