Jensen Huang spent 12.9 billion to buy the GitHub of the AI world, then turned around and sold out a robot duck
NVIDIA directly announced big news: it acquired Hugging Face, with a total price of about 12.9 billion US dollars, plus up to 1 billion US dollars in employee retention plans, bringing the total to nearly 13 billion!
A year ago, Hugging Face had rejected Lao Huang’s 500 million US dollars investment, fearing that the entry of a major shareholder would undermine the platform’s neutrality. As a result, nine months later, the entire company was taken over by Lao Huang.
What’s even more interesting is that at the same time this sky-high acquisition was announced, NVIDIA also turned around and sold out a robot duck called Microduck. In the first 24 hours of launch, it raked in 2.6 million US dollars, and at peak times, one unit was sold every 4 seconds on average. Now, if you place an order, you have to wait 4-6 months to receive it.
This duck looks cute and silly, but its specs are anything but ordinary
It stands 25 cm tall, weighs 800 g, and is equipped with 15 motors, a camera, a depth sensor, and dual IMUs. Its beak can pick up small objects, it can walk, play soccer, and even right itself if knocked over. It can even be modified to skate.
What’s more, it runs on the RK3566 chip, and the entire hardware stack, SDK simulation, and reinforcement learning code are all open source. Ordinary students and hobbyists can use it for embodied intelligence experiments, and it costs only 399 US dollars.
Hugging Face’s annual revenue is only 150 million US dollars. What’s expensive is the resources it holds: millions of open-source models and tens of millions of developers — this is the entry point to the AI world. Closed-source giants are all developing their own chips. Lao Huang wants to lock in the next source of computing power, use this low-cost robot duck to attract a large number of developers, and seize the physical AI ecosystem.
Honestly, I really want one
NVIDIA directly announced big news: it acquired Hugging Face, with a total price of about 12.9 billion US dollars, plus up to 1 billion US dollars in employee retention plans, bringing the total to nearly 13 billion!
A year ago, Hugging Face had rejected Lao Huang’s 500 million US dollars investment, fearing that the entry of a major shareholder would undermine the platform’s neutrality. As a result, nine months later, the entire company was taken over by Lao Huang.
What’s even more interesting is that at the same time this sky-high acquisition was announced, NVIDIA also turned around and sold out a robot duck called Microduck. In the first 24 hours of launch, it raked in 2.6 million US dollars, and at peak times, one unit was sold every 4 seconds on average. Now, if you place an order, you have to wait 4-6 months to receive it.
This duck looks cute and silly, but its specs are anything but ordinary
It stands 25 cm tall, weighs 800 g, and is equipped with 15 motors, a camera, a depth sensor, and dual IMUs. Its beak can pick up small objects, it can walk, play soccer, and even right itself if knocked over. It can even be modified to skate.
What’s more, it runs on the RK3566 chip, and the entire hardware stack, SDK simulation, and reinforcement learning code are all open source. Ordinary students and hobbyists can use it for embodied intelligence experiments, and it costs only 399 US dollars.
Hugging Face’s annual revenue is only 150 million US dollars. What’s expensive is the resources it holds: millions of open-source models and tens of millions of developers — this is the entry point to the AI world. Closed-source giants are all developing their own chips. Lao Huang wants to lock in the next source of computing power, use this low-cost robot duck to attract a large number of developers, and seize the physical AI ecosystem.
Honestly, I really want one