Giga AI

2wks agoupdate 237 0 0

Focusing on the development and commercialization of embodied intelligence (physical AGI) technology driven by the world model at its core, we create a full-stack solution for general-purpose robots from data generation, end-to-end decision-making to native ontology landing.

Location:
China
Language:
zh
Collection time:
2026-04-17
极佳科技Giga AI

Company Profile

Founded in January 2023, Giga Technology (full name “Beijing Giga Vision Technology Co., Ltd.”) is a company that specializes inAI macromodeltogether withembodied intelligence(AGI), headquartered in Changping District, Beijing, is a high-tech enterprise with the core objective of “building general artificial intelligence in the physical world”. With the core goal of “creating general artificial intelligence in the physical world”, the company provides software and hardware integrated solutions for physical intelligent robots by focusing on the trinity layout of "base model - ontology - scenario".2026 In March, the company completed the Pre-B round of financing of nearly 1 billion yuan, with investors including Semiconductor Jurassic Resources, CIC Capital, Shanghai Semiconductor Industry Fund and other industrial capital and state-owned platforms. In March, 2026, the company completed nearly 1 billion yuan of Pre-B round of financing, the investors cover SMIC Joyuan, CIC Capital, Shanghai Semiconductor Industry Investment Fund and other industrial capital and state-owned platforms, the valuation of the industry among the forefront.

Products & Services

The product line of GigaTech covers full-stack hardware and software solutions for embodied intelligence, including:

  1. Body-borne base molds (GigaBrain series)
    • As the “brain” of a general-purpose robot, it supports multi-sensor inputs such as image, point cloud, text, and body state, and outputs structured task planning and motion planning.
    • The open-source model GigaBrain-0.1 won the first place in the world in the authoritative large-scale real-machine evaluation, and GigaBrain-0.5m is the world's first reinforcement learning embodied base model based on the world model, which realizes high-efficiency learning and self-evolution.
  2. World Modeling Platform (GigaWorld series)
    • Built specifically for Vision-Language-Action (VLA) model training, it alleviates the data dilemma of existing models by generating highly generalizable embodied interaction data.
    • Releasing and open-sourcing GigaWorld-0 in December 2025 to support breakthroughs in real-time, training efficiency, and success rates for the World-Action Model.
  3. Native Ontology (Maker Series)
    • Full-stack self-developed robot hardware, the first humanoid robot Maker H01 released in November 2025, with more than 20 degrees of freedom, supporting fine manipulation and supple motion control, and has been opened for large-scale mass production delivery.
  4. Commercialization Scenario Solutions
    • Provides general-purpose robot solutions for automotive manufacturing, 3C electronics, warehousing and logistics, advanced tour guides, and home services.
    • In the field of automatic driving, the world model DriveDreamer has entered the programs of 20 mainstream car companies and automatic driving companies in China.

core technology

GigaTech's core technology centers around the “World Model”, building a closed-loop ecosystem from data generation to end-to-end decision-making:

  1. World Model
    • Combining Internet data, simulation data, teleoperation data, and real acquisition data to train an interactive physical world simulator to provide high-quality data support for embodied intelligence.
    • Technological breakthroughs include 4D driving scene reconstruction, video generation and fusion of world models, which significantly improves the generalization of autonomous driving and robotics tasks.
  2. End-to-End Big Model Architecture
    • Abandon the traditional sub-module design and adopt the end-to-end paradigm of “Video-in Action-out” (Video-in Action-out) to realize the optimization of the whole link from perception to decision-making.
    • The GigaBrain family of models supports long sequences of complex operations with enhanced generalization and real-time performance.
  3. Data-efficient training platform
    • Data Amplification (Data Amplification) is realized through the GigaWorld platform, which reduces the cost of real data collection and improves training efficiency by 10-100 times.

major client

GigaTech's clients cover a number of high-value scenarios, including:

  1. Automotive: 20 mainstream car companies such as Ideal Automobile, BYD, Xiaomi and Xiaopeng, as well as autonomous driving solution providers such as Horizon and Yigatong.
  2. Industry & Logistics: 3C electronics manufacturers and warehouse logistics operators, deploying robots for precision assembly and intelligent sorting.
  3. Service Areas: Advanced guided tour robots, home service robots, used in shopping malls, hotels and home scenarios.
  4. research organization: Collaborate with Tsinghua University, Institute of Automation of Chinese Academy of Sciences, and Technical University of Munich to promote technology iteration and scenario validation.

development prospect

By virtue of its technology leadership and commercialization landing speed, GigaTech is regarded as a head startup in the field of body intelligence, with great potential for future development:

  1. Accelerated technology iteration
    • Continuously optimize the world model and GigaBrain series to promote the penetration of general-purpose robots into more complex scenarios, such as healthcare and education.
    • Exploring the integration of the World-Action Model and the Large Language Model (LLM) for more natural human-computer interaction.
  2. Commercialization Scenario Expansion
    • We plan to release more native native robots adapted to different scenarios in 2026, and sprint to 1,000 unit deliveries for the whole year.
    • Deepen the cooperation with car companies and industrial enterprises to build a closed loop of the whole chain of data-training-testing and enhance customer stickiness.
  3. Capital and ecological support
    • Nearly 1 billion yuan of Pre-B round of financing for technology development and mass production delivery to provide financial security, investors covering industrial capital and state-owned platforms, help resources integration.
    • The open-source GigaWorld platform attracts developers to build an ecosystem and lower the technical threshold of the industry.
  4. Industry Trend Fit
    • Body intelligence is regarded as the key technology of the “fourth industrial revolution”, and the market size is expected to exceed one trillion yuan in 2026. As a technology pioneer, GigaTech is expected to lead the industry outbreak.

With the world model as its core, GigaTech has achieved technological breakthroughs and commercialization in the fields of autonomous driving and embodied intelligence through integrated hardware and software solutions and end-to-end large model architecture. With the expansion of scenarios and ecological improvement, the company is expected to become a leader in general artificial intelligence in the physical world and promote robots into thousands of industries.

data statistics

Related Navigation

No comments

none
No comments...