To make humanoid robots practical, they must be trained using real-world data. This week, humanoid robot maker Figure AI Inc. partnered with Brookfield Corp., a top asset manager with over $1 trillion in assets and 100,000 homes.
Brookfield, headquartered in Toronto, will assist Figure AI:
- Create a comprehensive and varied real-world dataset for humanoid pretraining
- Develop AI infrastructure to scale Helix, Figure’s custom vision-language-action (VLA) model
- Support the rollout of humanoid robots across new commercial environments
Figure’s Valuation Boosted as Brookfield Joins Major Series C Round
Brookfield also participated in Figure’s Series C funding round as part of the broader collaboration. Announced earlier this week, the round raised over $1 billion, valuing the San Jose-based company at $39 billion.
“This partnership marks major progress in our mission to build general-purpose humanoid robots,” said Figure AI CEO Brett Adcock. “Brookfield’s reach gives us a unique chance to gather real-world data on humanlike navigation and tasks—crucial for advancing general-purpose humanoids.”
In December 2024, Figure supplied its Figure 02 humanoid robots to a paying customer. The company received a 2024 RBR50 Robotics Innovation Award in recognition of the rapid progress in developing its humanoid technology.
Brookfield will Assist Figure in Developing its Dataset
Brookfield oversees more than 500 million square feet (46.4 million square meters) of commercial office space and 160 million square feet (14.8 million square meters) of logistics facilities. According to Figure, these settings are key to its data collection efforts for Helix.
By capturing human activity on video across various Brookfield locations, Figure aims to gather essential AI training data to help Helix teach its humanoid robots how to navigate, perceive, and operate in a wide range of human-centered environments. Figure noted that data collection is already underway in Brookfield properties and plans are in place to expand the initiative in the coming months.
Additionally, the partnership will investigate wider infrastructure collaboration, such as backing next-generation GPU data centers, providing real estate for robotic training facilities, and developing commercial applications throughout Brookfield’s global operations.
Brookfield and Figure will also consider opportunities to gradually introduce humanoid robots within Brookfield’s portfolio. Figure has begun initial commercial deployments with select customers, and integrating use cases within Brookfield’s platform highlights the potential to scale robot usage across industries.
Discover More About Humanoid Robots at RoboBusiness 2025
Experts say scaling humanoid robots requires addressing safety, proving advantages over current methods, and meeting uncertain market demand.
These and other topics will be explored at RoboBusiness 2025, October 15–16 at the Santa Clara Convention Center. The event will feature sessions on humanoids, physical AI, enabling technologies, design and development, business, and field robotics.
Keynote speakers include Jim Fan, NVIDIA’s director of AI and distinguished scientist, and Pras Velagapudi, CTO of Agility Robotics, who will present “Lessons Learned From the First Humanoid Deployments.”
The humanoid track at RoboBusiness will feature sessions such as:
- “Humanoids for Real Applications: Mastering Safety and Performance,” presented by Nikolai Ensslen, CEO of Synapticon
- “Advancements in Humanoid Actuation,” led by Jordan Schaeffler, strategic business development engineer at Novanta
- “Integrating Behavioral Science into Humanoid Design,” with Ram Devarajulu, vice president and head of robotics for North America at Cambridge Consultants
RoboBusiness, the premier event for commercial robotics, will feature 60+ speakers, a startup workshop, and the Pitchfire competition. Attendees will also have access to many networking opportunities and more than 100 exhibitors on the show floor.
Read the original Article on: The Robot Report
Read more: OpenMind Releases OM1 Beta, a Robot-Agnostic Open-Source OS