- Mar 19
- 5 min read

Tesla Bot Competitor's $675 Million Raise With Open AI (Figure AI)

Figure AI Secures $675 Million Funding and Partners with Open AI for Next Generation AI Models Development

Questions to inspire discussion:

❓ What is the amount Figure AI raised in their recent funding round?

$675 million

🤨 How did Figure AI demonstrate their robot's ability to operate a coffee machine autonomously?

Figure AI demonstrated their robot's ability to operate a coffee machine autonomously by showing off their bot autonomously operating a coffee machine.

💬 How will the recent funding accelerate Figure AI's commercial timeline?

The recent funding will allow Figure AI to stay in operation for quite a while and potentially start generating money before running out of funds.

🤨 How is OpenAI expected to help Figure AI in the future?

OpenAI is expected to help Figure AI in the future by allowing them to utilize the open language model to give commands, engage in end-to-end development, and enhance their algorithms through reinforcement learning. This partnership with OpenAI is anticipated to make Figure AI more robust and advanced in the field of AI development.

💬 Who are some of the notable investors and partners mentioned in the conversation?

Jeff Bezos is mentioned as an investor, but there is no confirmation about Amazon's involvement. Arc and Brett Winton are also discussed as potential investors, but their presence is not confirmed. Microsoft is highlighted as a significant partner in the conversation.

🤨 What role do Microsoft and Nvidia play in the partnership with Figure AI?

Microsoft provides AI infrastructure training and storage, while Nvidia supplies the chips for creating data warehouses.

❓ How does the processing speed of data warehouses and execution impact the bot's response time?

The processing speed of data warehouses and execution directly impacts the bot's response time. At lower levels, the core processes images quickly, while higher levels like planning and execution may have slightly slower speeds. This can result in a quick interpretation of the world at low levels, but a longer response time from the bot overall.

Key Highlights:

Figure AI secures significant funding and collaboration with Open AI for advanced AI models.
Accelerated commercial development timeline with increased funding for AI integration.
Importance of general education before specialization in AI, akin to doctors' training.
Partnership between Open AI and Microsoft for AI infrastructure training and storage, with involvement from Jeff Bezos and Amazon.
High-speed processing at low levels for quick response, but delays in full bot execution.
Significant advancements in robot walking capabilities through integration of neural net policy with body movements.
Understanding innate physics in predicting object behavior without explicit calculations.
Importance of real-world training for robots to enhance motor actions and understand friction.
Agility Robotics secures funding for robotics project in retail space with potential training methodology.

Clips:

00:00⚡️ AI humanoid robot company raises $675 million, partners with Open AI to accelerate development of Next Generation AI models.

Figure AI raised $675 million and is now valued at $2.6 billion.
Partnership with Open AI to develop Next Generation AI models.
Figure's robots gaining ability to understand language and take action.
Figure showcased autonomous operation of a coffee machine by their bot.
Impressive list of investors and partners including Microsoft and Open AI.
Partnership with Nvidia, Jeff Bezos, and Intel to accelerate commercial timeline.
Funding to be used for AI training, manufacturing, deploying more robots, and expanding engineering efforts.

04:30💡 Revolutionary advancements in AI training and development, paving the way for embodied AI and motion generation.

Implementation of open language model for end-to-end development
Transition towards visual learning models and embodied AI by Open AI
Utilizing neural nets to understand motion and physics for motion generation
Importance of generalized training over specific training for AI progress

08:04💡 Insights on recent AI developments, including funding, partnerships, and potential investors.

AI advancements in multiple areas of Athletics training for improved performance
Importance of broadening tasks for enhanced speed and dexterity
Focus on real-world AI applications like driving and movement
Necessity for AI to understand and execute movements for intelligent concepts
Potential improvements in AI software for smoother movements
Notable investors in AI initiatives and partnerships with Microsoft
Unconfirmed investors and missing names from previous reports
Speculation around Amazon's involvement in AI developments

11:48💡 Partnership between OpenAI, Microsoft, and Nvidia to provide AI infrastructure and training for humanoid robots.

Collaboration between OpenAI, Microsoft, and Nvidia for AI infrastructure and training.
Nvidia provides chips for data warehouses, Microsoft offers storage, and OpenAI contributes intelligence.
Possible cross-licensing and investment in kind between the organizations.
Potential development of specialized inference chips for embodied AI by Nvidia.
End-to-end neural net demonstrated for autonomous actions by the humanoid robot.
Low latency and low power consumption required for inference compute in the end product.
Quick response and latency at different levels of processing for interpreting the environment.
High H rates at low levels but slower refresh rate for overall response in robot controllers.

15:41 ⚙️ Overview of advancements in training methods for end-to-end neural network AI and Tesla's progress with their humanoid robot.

Training methods include simulations, real-world testing, and demonstration learning.
Tesla's AI team has made progress in achieving smoother movements and autonomy.
Initial walking was heuristic-based, but now likely integrated with neural net policy.
Recent updates show improvements in walking capabilities of the Tesla bot.

19:37 🤖 Advancements in Tesla's AI neural network and speculation on its capabilities and potential impact.

Tesla's AI neural network is progressing rapidly, with a walking speed of 6 meters per second.
Speculation suggests Tesla's AI may be creating video recreations to predict future scenarios while driving.
Elon Musk hinted at Tesla's AI understanding physics and making heuristic calculations.
The system is expected to infer object persistence and make real-time decisions similar to human drivers.

22:46💡 Challenges in training AI bots to understand real-world physics and movements for optimal performance.

AI bots need real-world experience to grasp physics intuitively
Tesla's Vision AI understands physics of movement better than Open AI
Training AI bots with diverse tasks improves their real-world capabilities
Simulation training provides limited understanding of real-world physics
Real-world friction and motor actions are crucial for AI bot fine-tuning
Increasing the number of AI bot copies aids in model refinement
Iterating and creating multiple copies of AI bots enhances their learning
Iterative process necessary for optimizing AI bot performance

26:26 ⚙️ Development progress and partnerships of humanoid robot for industrial tasks.

Focus on perfecting design before moving forward to avoid wasting time
Concerns about overproduction and need for modifications
Plans for completing humanoid robot hardware within six months
Partnership with BMW for commercial deployment in manufacturing tasks
Identification of tasks like sorting, picking, and logistics for robot deployment
Expectation of results by end of the year and completion in 12-24 months
Focus on training methodology and task iteration for faster progress