Robotics AI software company Skild AI raises $1.4 billion in funding

Today, U.S. robotics startupSkild AIAnnouncing the completion of approximately1.4 billionUSD (approximately Rs. 9.76 billion) Series CfinancingThe valuation rose to over14 billionU.S. dollars (about 97.63 billion yuan), reaching three times the level of seven months ago.

current round of financingLed by Softbank(math.) genusNVIDIA, Macquarie, 1789 Capital, Bezos family office Bezos Expeditions followed by Samsung, LG, Salesforceand other companies are also involved.

Deepak Pathak, co-founder and CEO of Skild AI, said the funds will be used for the rollout of Skild Brain software, training optimization and robot deployment expansion.

Skild AI sinceMay 2023Since its inception, it has raised a total ofAbout 2 billionThe U.S. dollar (about 13.94 billion yuan), is trying to build a “universal robot brain” to promote robots like humans by watching videos and learning by doing to complete tasks.

A startup team from Meta wants to make a “universal brain” that any robot can use.”

Skild AI was co-founded by Deepak Pathak and Abhinav Gupta.The duo worked together at Meta Robotics Labs in their early years., and later as a professor at the Robotics Institute at Carnegie Mellon University, respectively.

▲Skild AI co-founder and CEO Deepak Pathak (left) and co-founder and president Abhinav Gupta (right), Photo credit: Linkin

Back in 2022, the pair realized that robotics research was facing a bottleneck in getting reality off the ground: video demonstrations looked cool, but lacked the capability of a stable, replicable system.

As a result, they decided to turn to entrepreneurship with the goal ofBuilding a “Universal Robot Brain” that can run on standard GPUs without a custom chip”, achieving generalized execution capabilities across platforms, tasks, and scenarios.

Skild Brain built by Skild AI teamTraining based on large-scale human video and simulation exercisesThe system continuously optimizes the control strategy by generating a closed loop of feedback from action execution and error data. The system integrates the robot's internal sensing (e.g., joint motion, force sensing) and external visual signals, so that it can synchronize the understanding of its own state and environmental changes, and has strong task robustness.

Co-founder Gupta says that even if a robotic arm is damaged, it can still continue to accomplish tasks, “It's a security capability that robots haven't had before.".

Skild Brain currentlyAdapted to multiple robot forms, from quadrupedal platforms to robotic arms and humanoid robots can be deployed and can perform tasks in complex environments such as homes, warehouses, hospitals and construction sites.

▲Photo credit: Business Wire

In February 2024, the Skild AI team acquired a Yuuki humanoid robot and deployed Skild Brain into actual hardware for the first time, achieving perceptible experimental results in just one day.

Second, aiming at millions of job gaps, let the robot into the real work scene

As of 2025, Skild AI has servedMore than 8customers, revenues jumped from zero to severalten million dollars. Currently, the size of its team hasOver 100, with most of the members coming from tech companies such as Meta, Tesla, Google, Amazon and NVIDIA.

LG CNS has announced that it will work with it to develop humanoid robotics solutions, and Skild AI was also unveiled at NVIDIA's GTC conference in October 2025, where its software will be used to automate deployments at NVIDIA's GPU factory in Houston.

The Skild AI team believes that as robotic systems are progressively adapted to real-world needs, they have great potential for dangerous, repetitive, or undesirable positions, and are expected to fill more than a million vacant jobs in the United States.

Pathak also mentioned that “robots can eventually do almost all the work that humans can do, but that's not a threat, and we have plenty of time to prepare.”

Third, the robotics track heats up, attracting tech giants to raise stakes

Skild AI's high valuation and overfunding reflects the current change in the investment pattern of tech giants in the direction of robotics: no longer relying on asset-heavy mergers and acquisitions, but rather tending to bet on potential players through early-stage investments.

Samsung's investment in Skild AI is seen as a “lightweight layout” that allows it to keep an eye on the startup's technological developments and talent pool without incurring high acquisition costs. NVIDIA, on the other hand, has invested in companies such as Figure AI and Serve Robotics in addition to Skild AI.

Physical Intelligence, another Samsung-backed robotics AI company, has been valued at $5.6 billion (about Rs 39.05 billion) by the end of 2025 and is on a similar track to Skild.

At the same time, players in this space generally face similar challenges, especially in the deployment of consumer-oriented humanoid robots, where range, mobility and safety remain the sticking points. Even Tesla, with its long-launched Optimus program, is still in the attack phase.

Conclusion: the battle for the universal brain has just begun

Behind Skild AI's high valuation financing is the collective sprint of the robotics industry from showy demos to scale and landing.

As tech giants like NVIDIA and Samsung accelerate their bets, a consensus is forming around the evolution of robotic intelligence.

For the industry as a whole, the real inflection point may come from those general-purpose systems that were the first to run through the closed loop of training, adaptation to deployment.

Source: Bloomberg, Skild AI