SoftBank and Nvidia Close $1.4Bil Deal to Anoint Skild AI as the “Universal Brain” for Robotics at $14Bil Valuation
initially reported Dec 10, 2025. Last Update Jan 14, 2026
SoftBank and Nvidia have officially closed a $1.4 billion investment in Skild AI, valuing the two-year-old robotics company at over $14 billion. According to reports from Bloomberg and TechCrunch, the deal has transitioned from advanced talks to a finalized landmark round, marking one of the largest capital injections in the history of the robotics sector.
This valuation effectively anoints Skild AI as the dominant horizontal platform in the race to build general-purpose robotics. The deal more than triples the $4.5 billion mark the company set in May, when SoftBank led a $500 million Series B. With this close, total funding for the East Liberty-based company—which is already backed by Jeff Bezos, Sequoia Capital, Coatue, and Lightspeed—exceeds $2.1 billion across three rounds in just eighteen months. The updated cap table also adds strategic weight with new participation from 1789 Capital, Macquarie Group, Samsung, and Salesforce.
This investment functions as the software counterpart to SoftBank’s October 8 acquisition of ABB’s robotics division for $5.375 billion. The ABB unit generated $2.3 billion in revenue last year and employs 7,000 people manufacturing industrial robots for automotive and electronics factories globally. The valuation discrepancy underscores SoftBank’s thesis on value accrual, as they paid roughly 2.3 times revenue for ABB’s industrial hardware but are paying a significant premium for Skild’s intelligence layer, which has already scaled from zero to a $30 million revenue run rate in its first year of commercialization.
SoftBank CEO Masayoshi Son has effectively acquired the body and the brain separately. With a portfolio that already includes varying stakes in AutoStore, Berkshire Grey, and Agile Robots, SoftBank is positioning Skild to serve as the unifying foundation model across its entire hardware ecosystem. In the ABB acquisition announcement, Son explicitly stated his goal to fuse Artificial Super Intelligence (ASI) and robotics. Skild is positioned to serve as that ASI layer, deploying its operating system across ABB’s massive installed base once the acquisition closes in 2026.
Capital is flooding the robotics sector to answer a fundamental question regarding how the stack will mature. Competitors pursuing a vertical thesis, such as Figure AI (valued at $39 billion) and 1X (reportedly raising at $10 billion), are building the full stack by owning the robot, the brain, and the manufacturing. Conversely, Skild AI and Physical Intelligence (valued at $5.6 billion) represent the horizontal thesis, building a universal operating system that abstracts away the hardware layer so others can manufacture the metal. At $14 billion, Skild becomes the market’s dominant horizontal player, operating on the premise that robot hardware will commoditize while value concentrates in the intelligence layer that controls it.
Skild’s valuation is driven by its “Skild Brain” architecture, but recent strategic partnerships and technical breakthroughs have validated its ability to execute in enterprise environments. In addition to its 2025 agreements with LG CNS and Hewlett Packard Enterprise, the company has revealed a capability called “In-Context Learning.” This allows robots to adapt to new environments or mechanical failures—like a jammed motor or broken limb—in real time without needing to be retrained or updated.
Founders Deepak Pathak and Abhinav Gupta built the model on two pillars that bypass the need for hardware-specific data. The first is massive simulation, where Skild runs 100,000 simultaneous robot instances inside Nvidia’s Isaac Lab to generate a millennium of experience every few days. The second is internet video analysis at an unprecedented scale: training on 1,000 times more data than traditional models by extracting manipulation knowledge from millions of public web videos. Skild claims this approach reduces deployment costs to between $4,000 and $15,000 per robot, compared to $250,000 for conventional programming.
With this $1.4 billion infusion, Skild AI has secured the computational power and global reach necessary to move from R&D into the massive industrial deployments planned for later this year.