The next NVIDIA in robotics will likely emerge from the edge AI hardware space, not from companies building robots themselves. Companies like Wayve, which just closed a $1.2 billion Series D in February 2026 with backing from NVIDIA, Microsoft, Uber, and multiple automotive manufacturers, represent a critical inflection point: the infrastructure layer for autonomous systems is consolidating around edge computing, and whoever owns the chips and software stack that power on-device intelligence will define the economics of robotics deployment for the next decade. Unlike NVIDIA’s dominance in data center GPUs, the next infrastructure champion will win by solving the inverse problem—bringing AI inference closer to the edge, reducing latency, cutting connectivity costs, and enabling real-time decision-making in warehouses, vehicles, and factories.
The robotics market is already voting with its capital. In Q1 2026 alone, robotics startups secured over $2.26 billion in funding, with 70% flowing to warehouse and industrial automation companies. These aren’t training large AI models in the cloud; they’re deploying pre-trained models onto edge hardware where they can operate without constant internet connectivity. This shift mirrors the transition from mainframe computing to distributed systems—the winner won’t be the company with the biggest data center, but the company that controls the hardware and software architecture that makes distributed robotics intelligence economical and reliable.
Table of Contents
- Why Is Edge AI Infrastructure Becoming More Valuable Than the Robots Themselves?
- The Massive Funding Surge in Edge AI Hardware Companies
- The Autonomous Driving Connection and Broader Robotics Application
- NVIDIA Jetson Thor vs. Competing Hardware: The Specification War That Doesn’t Matter
- The Silent Killer: Software Integration Across Incompatible Platforms
- The Humanoid and Mobile Robotics Factor
- The Future Landscape and the Companies That Could Dominate
- Conclusion
Why Is Edge AI Infrastructure Becoming More Valuable Than the Robots Themselves?
Edge AI infrastructure is becoming the commodity in robotics because it’s the constraint that robotics companies can’t work around. A warehouse operator deploying 50 Exotec robots doesn’t care which company manufactured the robots—they care about reducing latency in object detection, cutting power consumption, and ensuring the system continues operating if the internet goes down. Exotec, Locus Robotics, Outrider, Covariant, and Path Robotics collectively raised over $1.5 billion in recent years, and almost all of them are now competing on the edge AI stack they deploy, not just the mechanical platform. The company that can offer a superior edge inference chip and software framework becomes the enabler for all of these robotics companies.
nvidia has recognized this dynamic and responded strategically. Rather than directly competing with edge AI startups, NVIDIA has co-invested in at least five companies (Wayve, Nuro, Figure, Bright Machines, and Skydio), positioning itself as a strategic co-investor to secure future inference workloads. This is a chess move, not defensive. NVIDIA knows that by partnering with the founders of edge robotics companies, it can influence their hardware roadmap and ensure NVIDIA chips remain the default choice. However, this also signals vulnerability—NVIDIA’s dominance in data center GPUs doesn’t automatically extend to edge devices, where power efficiency, form factor, and software integration matter more than raw compute.
The Massive Funding Surge in Edge AI Hardware Companies
Edge AI companies that closed funding rounds in 2025 and 2026 averaged $310 million per round—more than double the $140 million average for companies whose last round closed in 2023 or earlier. This 2.2x increase in funding size reflects how much money is chasing edge infrastructure bets. The global edge computing market is projected to grow at 33.0% CAGR from 2025 through 2033, reaching $327.8 billion by 2033. To put that in perspective, that’s faster growth than the AI training market and with fundamentally different economic characteristics—edge devices can’t be shared across customers, so there’s no marginal cost advantage to scale, which means the winner takes most market share, not most profit.
But there’s a critical limitation most investors and founders aren’t fully pricing in: edge AI infrastructure only becomes valuable at scale. A $310 million funding round sounds impressive until you realize that manufacturing, certifying, and deploying edge AI chips across thousands of customer deployments is capital-intensive. SambaNova’s SN50 chip claims 5x faster max speed than competitors and 3x lower total cost of ownership compared to GPUs for agentic AI workloads, but having superior specifications doesn’t guarantee market adoption. The hardware has to be paired with a software stack that robotics companies can actually integrate into their systems without hiring 50 new engineers. Wayve understands this—its $60 million Series D extension in April 2026 from AMD, Arm, and Qualcomm Ventures wasn’t just about hardware; it was about securing partners who could help distribute that hardware across multiple industries.
The Autonomous Driving Connection and Broader Robotics Application
Wayve’s trajectory deserves close attention because it reveals how edge AI infrastructure companies achieve scale. Wayve raised its initial $1.2 billion from not just tech investors but from Uber, Mercedes-Benz, Nissan, and Stellantis—every partner was a potential customer for edge AI infrastructure. Autonomous driving is a particularly brutal test case for edge AI because the cost of network latency is literally catastrophic. A self-driving car can’t afford a 500ms round trip to the cloud to make a lane change decision.
this reality forces autonomous driving companies to solve the hardest version of the edge AI problem, and once they’ve solved it, that solution is transferable to warehouse robots, delivery drones, and manufacturing systems. The funding patterns show that warehouse and logistics robotics are the most immediately viable market for edge AI infrastructure. Companies like Exotec have already deployed thousands of systems that need real-time computer vision and path planning. But there’s a timing mismatch: humanoid robotics (Figure raised $1.7 billion across three rounds) is attracting more headlines, but warehouse robotics is generating more revenue and more consistent edge AI infrastructure demand. The next NVIDIA will likely win in warehouse automation first, then expand to other domains, rather than waiting for humanoid robots to become mainstream.
NVIDIA Jetson Thor vs. Competing Hardware: The Specification War That Doesn’t Matter
NVIDIA has launched its Jetson Thor processor as the successor to the Jetson Orin, offering 7.5x more AI compute and 3.5x greater energy efficiency. Amazon Robotics, Boston Dynamics, Figure, and Caterpillar have already adopted it. On paper, this dominance seems insurmountable. But comparing chip specifications misses the real competition.
The war isn’t between NVIDIA and SambaNova; it’s between hardware vendors who can deliver complete integrated solutions (chip + software framework + customer support) and hardware vendors who sell chips and hope customers figure out the rest. Wayve’s edge AI advantage isn’t just the chip; it’s the combination of the chip, the software stack, and the deep integration with customer deployments. When Mercedes-Benz agrees to use Wayve’s technology, they’re committing to Wayve’s entire stack, not just swapping in a different processor. This is the tradeoff that hardware-centric companies like SambaNova and traditional GPU vendors face: they can win the specification sheet, but they lose the integration layer where the real value lives. The company that owns both will achieve NVIDIA-like margins and market dominance.
The Silent Killer: Software Integration Across Incompatible Platforms
Here’s the limitation that most edge AI companies aren’t talking about openly: deploying edge AI inference across heterogeneous hardware is vastly more complex than training models in the cloud. A warehouse operator might have NVIDIA Jetson hardware in some robots, but also need to support AMD, Qualcomm, and custom SoCs from smaller vendors. Building a software framework that can transparently move between these platforms without retraining models or rewriting code is exponentially harder than it sounds. Wayve’s $60 million Series D extension from AMD, Arm, and Qualcomm Ventures explicitly addresses this challenge—by partnering with chip manufacturers, Wayve ensures its software stack is optimized for multiple hardware targets simultaneously.
The warning here is critical for investors and customers: any edge AI company claiming platform-agnostic inference is overselling. The companies that will dominate the market are those that pick a specific combination of chip (usually NVIDIA initially) and software stack, then make that combination so valuable that customers accept the lock-in. Once a customer has deployed 50 robots running NVIDIA Jetson with Wayve’s software stack, switching to a competing platform becomes prohibitively expensive. The next NVIDIA won’t emerge from a startup that promises to work with everyone; it will emerge from a startup that promises to work better with one company’s hardware, then expands from there.
The Humanoid and Mobile Robotics Factor
Figure AI’s $1.7 billion in funding over three rounds has created a narrative that humanoid robotics is the future of edge AI. But this narrative inverts the actual causality. Figure isn’t well-funded because edge AI is hard; it’s well-funded because investors believe humanoid robots will eventually drive massive edge AI infrastructure demand. The robots themselves are infrastructure—they’re the proof of concept that edge AI is valuable enough to justify the $300 million funding rounds now flowing to hardware companies.
The immediate edge AI market leader will likely come from the warehouse automation space, where the ROI is measurable today. Hellbender closed a $12.5 million seed round in May 2026 to manufacture physical AI infrastructure, including on-edge camera systems specifically designed for robotics. This $12.5 million seed round to a manufacturer sounds small compared to Wayve’s $1.2 billion, but it’s actually the signal that matters: foundational physical AI hardware is attracting dedicated manufacturers. Over the next 5-7 years, these manufacturers will consolidate, and the winner will have achieved NVIDIA-like dominance in edge robotics hardware.
The Future Landscape and the Companies That Could Dominate
The robotics industry is entering a phase where infrastructure beats product. The companies that will dominate robotics in 2030 won’t necessarily be the robotics companies themselves; they’ll be the infrastructure companies that enabled those robots to operate economically at scale. Wayve’s combination of $1.2 billion in funding, backing from automotive manufacturers and tech giants, and already-proven deployment of autonomous systems across multiple customers, positions it as a strong candidate for this role. But the competition will intensify rapidly.
What’s not yet resolved is which hardware layer will matter most. If edge inference hardware becomes commoditized quickly (similar to how GPU commodity pricing evolved), then software stack and integration become the only defensible moat. If edge chips remain specialized and difficult to optimize for, then hardware vendors control pricing and customer lock-in. The next NVIDIA will be the company that answers this question correctly first, then scales aggressively before competitors can react. For robotics companies, the strategic choice is clear: partner early with the edge AI infrastructure company you believe will win, because switching costs are prohibitively high once deployment begins at scale.
Conclusion
The case for an edge AI company becoming “the next NVIDIA in robotics” rests on three fundamental shifts: the migration of AI inference from cloud to edge, the massive consolidation of robotics funding around infrastructure layers, and the realization that hardware integration matters more than raw chip performance. Wayve’s $1.2 billion Series D and subsequent $60 million extension from AMD, Arm, and Qualcomm perfectly illustrate this dynamic—a company explicitly building edge AI infrastructure for autonomous systems is attracting more capital and strategic backing than companies building the robots themselves. For robotics companies, investors, and manufacturers watching this space, the critical question is no longer “which robot will win,” but “which edge AI infrastructure company will become the standard platform.” The answer will determine the margins, the competitive landscape, and the dominance hierarchy of robotics for the next decade.
The next NVIDIA is already raising capital. It’s just not obvious yet whether it will be Wayve, an NVIDIA co-investment, or a company still in stealth mode. What is clear is that whoever answers the edge AI integration problem at scale will own more of the robotics value chain than any robot manufacturer ever could.
