How to break the commercial dilemma?
Recently, Zhejiang Youlu Robotics Technology Co., Ltd. (hereinafter referred to as "Youlu Robotics") fired the "first shot" of embodied intelligence commercialization with over 100 million yuan in financing and over 10 million orders.
The emergence of ChatGPT has set off a new round of AI craze around the world. However, to achieve true "general artificial intelligence" (AGI), it must be rooted in the physical world and understand and control various objective entities. Therefore, the concept of embodied intelligence has become popular again, and large models are gradually sinking into reality.
Although Google's PaLM-E and Stanford's VoxPoser have made encouraging progress and can directly understand natural language commands and break them down into a series of actions, embodied intelligence still faces many obstacles that need to be overcome in its true industrialization.
Universal brain, decoding trillion-dollar market
It is in this context that Youlu Robot came into being.
The company's founder, Chen Junbo, holds a doctorate in computer science from Zhejiang University. He was previously the head of anti-cheating algorithms at Alibaba Cloud's search department, the head of algorithms at Taobao's data lab, the head of algorithms at Cainiao's ET lab, and the business head of Alibaba DAMO Academy's autonomous driving lab. Alibaba's delivery robot "Little Donkey" is the work of the DAMO team.
In 2023, Chen Junbo left home to start his own business and founded Youlu Robotics with former Alibaba Robotics CEO Gu Zulin and former Zhejiang Netxun Hengtian Software Vice President Wang Jian, focusing on the integrated innovation of large models and embodied intelligence.
In order to truly implement embodied intelligence, several hurdles must be overcome: first, the integration of large models with the real physical world, directly perceiving and manipulating objective entities through models; second, training models in massive real-scene data, and continuously iterating and optimizing; and third, seamlessly connecting advanced technologies with traditional manufacturing to meet the actual needs of production and application.
In order to train a truly usable embodied intelligence model, a large amount of data from the real physical world must be obtained, which must be based on commercialization. Therefore, Youlu Robotics has set its sights on providing a "universal intelligent brain" for various professional equipment and focusing on the generalization and upgrading of robot forms, which has a unique industrial advantage in China.
my country is a major manufacturing country, and its overall scale has ranked first in the world for 14 consecutive years. However, most manufacturing companies lack the accumulation of artificial intelligence technology and find it difficult to develop exclusive intelligent systems for their own products.
It is this huge market gap that Youlu Robot is targeting.
In the future, the overall market size of wheeled robots, composite robots and emerging humanoid robots is expected to reach trillions of yuan. The "universal brain + professional equipment" model advocated by Youlu can minimize the cost of intelligent transformation of enterprises, improve the input-output ratio, and pave the way for the intelligent upgrade of enterprises.
In line with the above ideas, Youlu Robotics and Alibaba Cloud Tongyi Qianwen jointly released the LPLM-10B (Large Physical Language Model) model in 2023, which solves the defects of traditional embodied models such as high latency, low accuracy, and poor generalization ability. It can extract a lot of common information from cross-modal, cross-scenario, and cross-industry data, and is highly versatile.
In March this year, Youlu Robotics launched the Master 2000 universal embodied intelligent brain based on LPLM-10B. This product can directly connect to traditional equipment, giving it intelligent capabilities and is plug-and-play.
With its outstanding performance, Youlu Robot soon won an order worth over 10 million yuan from a head cleaning equipment manufacturer, which is also the first high-value order in the field of embodied intelligence in the strict sense. It is reported that the product has started mass production and delivery.
Over 100 million yuan in financing to support commercialization
Recently, the company announced that it has completed a financing of over 100 million yuan, with investments from Innovation Works founded by Dr. Kai-Fu Lee, Yuanjing Capital founded by Alibaba co-founder Wu Yongming, and BV Baidu Ventures.
Moreover, investors’ valuation of Youlu Robot is not only based on short-term profits, but also focuses on the huge long-term development prospects of embodied intelligence.
According to Chen Junbo's vision, Youlu's mission is to equip every traditional device with a "universal intelligent brain" to make the machine "smarter". Relying on the universal advantages of LPLM technology, it can fully leverage trillions of dollars in business opportunities.
In addition, Youlu Robotics will also start to open up its own brand to help partners achieve brand intelligence upgrades. It is foreseeable that in the future, Youlu Robotics will deepen cooperation with more manufacturing partners to jointly build an upstream and downstream ecosystem of the embodied intelligent industry chain and accelerate the development of the entire industry.
Last words
From theory to practice, from laboratory to industrial-level application, if we want to make the big model truly land, conquering commercialization is the only way.
In fact, embodied intelligence is widely regarded as an important direction for the implementation of big model technology in the physical world. Compared with traditional machine learning methods, embodied intelligence based on big models has natural advantages in interactive understanding, multimodal perception, and cross-domain generalization, and is expected to become a key step towards AGI.
However, for embodied intelligence, the current technological breakthroughs have only gone halfway. The next decisive step is how to connect with traditional manufacturing and build a sustainable business model.
Youlu Robotics has pioneered the feasible path of "universal brain + professional equipment".
By giving the LPLM large model a universal cognitive structure and behavioral norms, it can be used as a basic model for various different forms of robots and seamlessly integrated with existing production lines, which will significantly reduce the cost of intelligent transformation and facilitate the rapid popularization of the product in a wider range of fields.
This "shared intelligence" model has a natural moat advantage.
Startups often have limited funds, making it difficult for them to simultaneously possess the dual capabilities of large-scale model research and development and industrial product manufacturing; and traditional manufacturing giants such as automobiles and construction machinery lack the genes for algorithm modeling, making it difficult for them to master large-scale model technology in the short term.
Youlu’s “combination punch” undoubtedly hit the mark.
If it can continue to maintain its technological leadership and commercialize in more vertical fields, Youlu Robotics may play a hub role in the entire embodied intelligence ecosystem and become an emerging force that cannot be ignored.
Previous article:Mitsubishi Electric invested in industrial robots
Next article:Siasun joins hands with Baidu, targeting this field!
- Popular Resources
- Popular amplifiers
- Using IMU to enhance robot positioning: a fundamental technology for accurate navigation
- Researchers develop self-learning robot that can clean washbasins like humans
- Universal Robots launches UR AI Accelerator to inject new AI power into collaborative robots
- The first batch of national standards for embodied intelligence of humanoid robots were released: divided into 4 levels according to limb movement, upper limb operation, etc.
- New chapter in payload: Universal Robots’ new generation UR20 and UR30 have upgraded performance
- Humanoid robots drive the demand for frameless torque motors, and manufacturers are actively deploying
- MiR Launches New Fleet Management Software MiR Fleet Enterprise, Setting New Standards in Scalability and Cybersecurity for Autonomous Mobile Robots
- Nidec Drive Technology produces harmonic reducers for the first time in China, growing together with the Chinese robotics industry
- DC motor driver chip, low voltage, high current, single full-bridge driver - Ruimeng MS31211
- Innolux's intelligent steer-by-wire solution makes cars smarter and safer
- 8051 MCU - Parity Check
- How to efficiently balance the sensitivity of tactile sensing interfaces
- What should I do if the servo motor shakes? What causes the servo motor to shake quickly?
- 【Brushless Motor】Analysis of three-phase BLDC motor and sharing of two popular development boards
- Midea Industrial Technology's subsidiaries Clou Electronics and Hekang New Energy jointly appeared at the Munich Battery Energy Storage Exhibition and Solar Energy Exhibition
- Guoxin Sichen | Application of ferroelectric memory PB85RS2MC in power battery management, with a capacity of 2M
- Analysis of common faults of frequency converter
- In a head-on competition with Qualcomm, what kind of cockpit products has Intel come up with?
- Dalian Rongke's all-vanadium liquid flow battery energy storage equipment industrialization project has entered the sprint stage before production
- Allegro MicroSystems Introduces Advanced Magnetic and Inductive Position Sensing Solutions at Electronica 2024
- Car key in the left hand, liveness detection radar in the right hand, UWB is imperative for cars!
- After a decade of rapid development, domestic CIS has entered the market
- Aegis Dagger Battery + Thor EM-i Super Hybrid, Geely New Energy has thrown out two "king bombs"
- A brief discussion on functional safety - fault, error, and failure
- In the smart car 2.0 cycle, these core industry chains are facing major opportunities!
- The United States and Japan are developing new batteries. CATL faces challenges? How should China's new energy battery industry respond?
- Murata launches high-precision 6-axis inertial sensor for automobiles
- Ford patents pre-charge alarm to help save costs and respond to emergencies
- New real-time microcontroller system from Texas Instruments enables smarter processing in automotive and industrial applications
- Silicon Labs EFM32PG22 Development Kit enables you to create whatever you want! Application in progress!
- Verilog state machine programming skills
- Review summary: Qinheng RISC-V core BLE 5.3 wireless MCU CH582
- MSP-EXP430F5529LP Development Board 002-GPIO Library Function
- Lederui specializes in Type-C interface and HOST solution for laptops/tablets
- Domestic FPGA Gaoyun GW1N-4 series development board evaluation - unboxing + hardware part 2
- Please help me take a look at this strange MICRO BIT board
- Free simulation software SimulIDE
- Many branches of MicroPython
- Choosing Machine Learning Processors at the Edge