Last Updated on May 3, 2026 by Eytan Bijaoui
⚡ Quick Answer: Robotics is the new AI for VCs. Q1 2026 saw $8.5B+ invested in robotics startups — more than all of 2024. Figure AI ($39B), Skild AI ($14B), and Neura Robotics lead a massive shift from software-only AI to physical intelligence.
📅 Last updated: March 29, 2026
Quick question. What’s the most valuable AI company that isn’t OpenAI or Anthropic?
It sits alongside the broader Q1 2026 venture funding patterns we tracked earlier this year.
If you said Databricks, Stripe, or some other software darling, I’d have agreed with you six months ago. But the answer today is Figure AI. A robot company. Valued at $39 billion after closing a $1 billion Series C that included NVIDIA, Salesforce, Intel Capital, and Qualcomm. A company that builds humanoid robots. Like, actual walking, moving, lifting-stuff robots.
And Figure isn’t even the wildest part of the story.
The Quarter That Robots Won
In the first quarter of 2026, robotics startups have raised over $8.5 billion. To put that in context, the entire global robotics venture market raised about $7 billion in all of 2024. We’ve already blown past a full year’s total, and it’s March.
The numbers are staggering when you line them up.
Skild AI, a Pittsburgh company building what they call a universal “brain” for robots, raised $1.4 billion at a $14 billion valuation. They tripled their valuation in seven months. Their entire revenue last year was $30 million. That’s a 466x revenue multiple. In a down market for SaaS companies trading at 8x.
Apptronik pulled in $520 million at a $5.5 billion valuation, backed by Google, Mercedes-Benz, John Deere, and the Qatar Investment Authority. Mind Robotics, a Rivian spinout that barely existed a few months ago, raised $500 million in a Series A. A Series A. Five hundred million dollars.
And Rhoda AI closed $450 million to build robots that learn from watching YouTube videos. Seriously. Their approach is to train robot behavior models on millions of hours of internet video, then drop that intelligence into physical machines.
All of this happened while the entire tech media was writing about chatbots, coding copilots, and whether AI is going to replace your accountant.
Why Robots, and Why Now
I’ve been covering AI’s impact on startups for a while now, and this caught me off guard. Not the fact that robotics is getting funded. But the scale and speed of the shift.
Three things happened at roughly the same time to make this possible.
Foundation models got physical. The same transformer architecture that powers ChatGPT turns out to work for robots too. Skild AI’s whole thesis is that you can pre-train a single model on video data and physics simulations, then deploy it across completely different robot types. One brain, many bodies. Six months ago, this was a research paper. Now it’s a $14 billion company.
Hardware got cheap. This is the less sexy explanation but probably the more important one. The cost of the sensors, actuators, and compute modules that go into a robot has dropped roughly 40% since 2023. Lidar that cost $10,000 three years ago costs $800 now. Custom AI chips from NVIDIA and Qualcomm are being designed specifically for edge robotics. The physical infrastructure for building robots at scale just crossed an affordability threshold.
The labor crisis got real. Manufacturing, logistics, agriculture, and healthcare are all facing chronic worker shortages that aren’t going away. The US alone has over 600,000 unfilled manufacturing jobs. Japan’s working-age population has shrunk by 8 million people since 2000. Robots aren’t replacing workers who want their jobs. They’re filling positions that nobody wants and nobody’s available for.
When a market opportunity is this obvious, this large, and this urgent, capital follows. That’s what we’re seeing.
The Numbers That Explain the Hype
Let me give you some context for why investors are pouring money into robots at rates that make the SaaS boom look quaint.
Global manufacturing output is roughly $16 trillion per year. Logistics and warehousing add another $5 trillion. Agriculture is $4 trillion. Healthcare services, which increasingly involve physical assistance and monitoring, sit around $9 trillion globally.
Add it all up and you’re looking at $30+ trillion in industries where the core work involves moving, building, sorting, picking, assembling, or assisting physical things in the physical world.
Software couldn’t really touch this. You can put a SaaS dashboard on top of a warehouse, but the actual work of picking items off shelves and packing them into boxes? That required human hands. Until now.
The bet that VCs are making is straightforward. If AI models can now understand physical environments, control robotic arms and legs, and learn new tasks from observation rather than explicit programming, then every industry that involves physical labor becomes an AI market. Not a software market. An AI-plus-hardware market.
We just covered how AI is disrupting the $5 trillion professional services market. But professional services is small compared to the physical economy. If robots get even 5% penetration into manufacturing, logistics, and agriculture in the next decade, that’s $1.5 trillion in addressable market. At 10%, it’s $3 trillion.
Figure AI is planning to ship 100,000 humanoid robots in the next four years. At any reasonable price point per unit, that’s a multi-billion-dollar revenue line. That’s why a robot company is worth $39 billion before it has meaningful revenue. Investors are pricing in the physical economy, not the software economy.
The Part Where This Isn’t Just About Giant Robots
Here’s where it gets actually relevant if you’re a founder without $500 million in the bank.
The humanoid robot companies are getting the headlines. Figure, Apptronik, Boston Dynamics. But the real opportunity for startups is in what I’d call the “picks and shovels” layer of the robotics revolution.
Every robot needs perception software (what am I looking at?). Every robot needs planning intelligence (what should I do next?). Every fleet of robots needs orchestration (which robot goes where?). Every warehouse deploying robots needs integration with their existing systems. Every factory adopting robots needs safety certification. Every logistics company using autonomous delivery needs regulatory compliance tooling.
None of this is as glamorous as building a humanoid that walks like a human. But all of it is necessary, all of it is underbuilt, and all of it can be done by a small team with deep domain knowledge.
I talked to a founder last month who left Amazon Robotics after 6 years. He’s building a tool that monitors robot fleet performance in real time, basically an observability platform for physical AI. Think Datadog, but for robots. No venture funding yet, just him and a co-founder, and three paid warehouse customers already using the beta. The one-person startup thesis applies here just as much as it does in software.
The comparison to early cloud computing keeps coming up, and I think it’s apt. When AWS launched in 2006, the headline was “Amazon is renting servers.” But the real startup gold rush wasn’t in building clouds. It was in building everything that runs on clouds. Monitoring tools, security layers, DevOps platforms, cost optimization, compliance frameworks. An entire ecosystem of multi-billion-dollar companies emerged to serve the infrastructure that Amazon built.
Robotics is at that same inflection point. The big players are building the “cloud” (the foundation models, the humanoids, the hardware platforms). The opportunity for founders is in building everything that makes those robots useful in specific industries.
Where the Gaps Are Actually Biggest
If I were starting a company in robotics right now (and believe me, the thought has crossed my mind even though I’m probably the least technical person you’ll meet), here’s where I’d look.
Robot-as-a-Service pricing and financing. Nobody has figured out how to sell robots to mid-market companies. A humanoid costs somewhere between $50,000 and $150,000 today. Most manufacturing companies can’t do a six-figure capex purchase for an unproven technology. But they CAN do a monthly subscription. Whoever builds the “robot leasing and management” platform, the Flexport of physical AI, is sitting on a gold mine. The SaaS model is dying for software, but it might be exactly what robotics needs.
Vertical-specific robot training. Skild AI is building the general-purpose brain. But a general-purpose brain isn’t enough for a food processing plant that needs a robot to debone a chicken at 90 units per hour with USDA compliance. The startups that take general foundation models and fine-tune them for specific industrial tasks will own the last mile of robotics deployment. This is the same pattern we saw with LLMs, where general models got good and then vertical-specific applications captured all the value.
Safety and compliance infrastructure. When a chatbot hallucinates, someone reads wrong information. When a robot malfunctions, someone could get hurt. The safety requirements for physical AI are orders of magnitude more complex than for software AI. Regulatory frameworks are just starting to emerge. The startup that builds the “trust and safety” layer for robotics, the certification, monitoring, incident reporting, and audit trail infrastructure, will become essential to every company deploying robots.
Human-robot workflow design. This one is underrated. Most factories and warehouses aren’t going to go fully autonomous overnight. They’re going to have humans and robots working side by side for years. Designing those hybrid workflows, figuring out handoffs, task allocation, exception handling when the robot can’t figure something out, that’s a specialized expertise that barely exists yet. It’s consulting meets software, and it’s going to be huge.
The Uncomfortable Parallel to the AI Bubble Conversation
I want to be straight with you about something. I look at some of these valuations and I get nervous.
Skild AI at a 466x revenue multiple. Figure AI at $39 billion before shipping at scale. Mind Robotics raising $500 million in a Series A for a company that’s months old. These numbers feel familiar. They feel like the kind of numbers we saw in AI software last year, the kind of numbers that led to the conversations about whether we’re in a bubble.
And maybe we are. Maybe some of these robotics companies will face the same reckoning that hit enterprise SaaS when the market repriced. Maybe the timeline from “cool demo” to “profitable at scale” is longer than VCs are pricing in. Maybe building and deploying physical machines in messy, regulated, real-world environments is harder than building software that runs in a clean cloud environment. (Spoiler: it definitely is.)
But here’s why I don’t think the parallel is exact. The industries that robots are targeting have a genuine, quantifiable, screaming need. Unlike the software job market where AI is redistributing work, manufacturing and logistics don’t have enough humans to do the work at all. That’s not a theoretical market. That’s 600,000 unfilled American manufacturing jobs today. That’s Japanese factories running at 70% capacity because they literally cannot find workers.
When the demand is real and urgent and growing, bubbles in specific company valuations matter less than the direction of the overall market. Some robotics companies will fail. Some valuations will correct. But the underlying shift toward physical AI is not reversing.
What This Means for Your Next Idea
If you’re a founder who has been laser-focused on software because that’s what you know, I get it. Software is comfortable. The margins are great. The distribution is digital. You ship an update and everyone gets it instantly.
But I want you to consider something. The wave of software disruption we’ve been tracking, the SaaSpocalypse, the AI agent revolution, the collapse of per-seat pricing, that wave is getting crowded. Every founder and their cousin is building an AI agent for something. The competition is brutal and the incumbents (Microsoft, Google, Anthropic, OpenAI) are moving fast.
Physical AI? It’s wide open. The incumbents don’t exist yet. The playbook hasn’t been written. The pricing models are still being invented. And the total addressable market is 5 to 10 times larger than enterprise software.
You don’t need to build a humanoid robot. You don’t need a PhD in mechanical engineering. You need to understand a specific physical workflow in a specific industry and figure out where AI-powered robotics can make that workflow 10x better. Then you need to validate that someone will pay for it before you build anything.
The biggest venture shift of 2026 isn’t happening in your browser. It’s happening in warehouses, on factory floors, in agricultural fields, and in operating rooms. The question isn’t whether robots are coming. The $8.5 billion in Q1 funding answered that.
The question is whether you’re going to keep building for the screen, or start building for the world.
