In partnership with
While the tech world debates whether flying drones will ever deliver your burrito, a fleet of sidewalk robots quietly completed their 9 millionth delivery.
Starship Technologies closed a $50 million Series C on October 15, bringing total funding to $280 million. The Estonian company operates 2,700+ robots across 270+ locations in seven countries. The robots make 2 autonomous road crossings per second.
Founded in 2014 by Skype co-founders Ahti Heinla and Janus Friis, Starship took a contrarian bet. Everyone else chased flying drones. Starship built six-wheeled coolers that travel on sidewalks at walking speed. The approach solved problems that still ground competitors: regulatory approval across seven countries, all-weather reliability, and actual profitability.
The business case works. Starship operates at positive gross margins while competitors burn cash on helicopter-sized delivery drones that regulators won't approve. The robots cost a fraction of human delivery, enabling sub-30-minute deliveries that would be uneconomical otherwise.
The company dominates two markets. European urban delivery. American college campuses. Now it's using the new capital to crack a third: US cities. Plural led the round, with participation from Karma.vc, Latitude, Coefficient Capital, and EU-backed SmartCap.
Starship has partnerships with Grubhub, Bolt, DoorDash's Wolt, and Delivery Hero's Foodora. While competitors chase the future, Starship became the delivery infrastructure apps already plug into.
The robots handle the boring middle. A restaurant loads the order. The robot navigates sidewalks autonomously. Customers unlock it with their phone. Nobody talks about the technology anymore. It just works.
Design lesson: Sometimes the unsexy solution wins. Flying drones make better demos. Sidewalk robots make better businesses.
The Einride Pod
Sweden Built a Truck With No Cab.
Then Drove It Across a Border.
Einride closed $100 million on October 1. The Swedish company then completed the world's first autonomous international border crossing from Sweden into Norway through the Ørje checkpoint.
The truck had no driver. No cab. No steering wheel. Just the Einride Pod, a purpose-built autonomous freight vehicle, carrying freight across an international border without human intervention.
This matters because Einride designed the vehicle from scratch for autonomy. Most autonomous trucking companies retrofit existing cabs. Remove the driver but keep the steering wheel, the seat, the entire human interface. Einride asked a better question: what if you never needed a cab?
The result is the Einride Pod. A purpose-built autonomous freight vehicle controlled by remote operators when needed and its "Einride Driver" software the rest of the time. The company operates one of the world's largest heavy-duty electric truck fleets for customers like PepsiCo, Carlsberg, and DP World across Europe, North America, and the UAE.
The border crossing proved the technology works in real-world complexity. Customs. Different regulations. International data exchange through Norway's Digitoll system. All handled autonomously.
Total funding now sits at $654 million. EQT Ventures led the round with participation from quantum computing company IonQ. The company is eyeing a US IPO.
Einride faces the same challenge every autonomous freight company does: the expensive crawl from development to commercial scale. Co-founder Robert Falck stepped aside as CEO five months ago. CFO Roozbeh Charli took over to focus on scaling operations.
The company runs three business lines. Electric big rigs for customers who want zero-emission freight now. Autonomous pods for fixed routes. Saga planning software for shippers who need better logistics. The strategy hedges bets while building toward full autonomy.
Why this matters: When everyone else tries to make autonomous trucks look like trucks, Einride rethought the entire form factor. That's either brilliant or foolish. The $654 million suggests investors think it's brilliant.
Figure 03
Figure Built a Factory. Now It Can Build 12,000 Humanoids a Year.
Figure launched Figure 03 on October 9. The third-generation humanoid isn't just better hardware. It's the first humanoid designed for manufacturing at scale.
Most humanoid robots are engineering prototypes. Custom parts. CNC machining. Assembly by hand. Each unit costs six figures and takes weeks to build. Figure 03 flips that model.
The company built BotQ, a dedicated manufacturing facility in the Bay Area. First-generation capacity: 12,000 robots per year. Four-year goal: 100,000 units total. Instead of outsourcing to contract manufacturers, Figure brought critical systems in-house to control quality and iteration speed.
The redesign cut costs dramatically. Figure 02 relied on expensive CNC machining. Figure 03 uses die-casting, injection molding, and stamping. Higher upfront tooling costs. Much lower per-unit economics as volume scales.
The hardware improvements matter. Twice the camera frame rate. One-quarter the latency. 60% wider field of view per camera. Each fingertip detects forces as small as three grams. That's sensitive enough to feel a paperclip.
The robot charges wirelessly. Step onto a charging pad. Pull 2 kilowatts. Walk away when done. It uploads terabytes of training data over 10 Gbps mmWave. The entire fleet learns from every deployment.
Figure 03 weighs 9% less than Figure 02. Soft foam and washable textiles instead of hard metal. The battery passed UN38.3 certification for home use. Wireless charging means near-continuous operation as long as the robot can dock periodically.
The commercial pitch is simple. Faster actuators for pick and place. Better cameras for navigation. More compliant hands for varied objects. Custom uniforms and branding. Near-continuous uptime with inductive charging.
Design lesson: You can't scale humanoids without manufacturing. Figure stopped building prototypes and started building a production line.
By The Numbers
9 million - Deliveries completed by Starship's sidewalk robot fleet
2/second - Autonomous road crossings by Starship robots globally
$654 million - Total capital raised by Einride since 2016
0 - Steering wheels in Einride's autonomous freight pods
12,000/year - Production capacity at Figure's BotQ manufacturing facility
3 grams - Minimum pressure detected by Figure 03's fingertip sensors
100,000 - Humanoid robots Figure plans to manufacture over four years
What I'm Watching
Can ground robots succeed where flying drones failed? Starship crossed 9 million deliveries. Zipline and Wing keep trying to make aerial delivery work. One operates at city scale. The others run limited pilots. Is this another VHS vs. Betamax, or can both models coexist?
Who builds the humanoid supply chain? Figure vertically integrated batteries, actuators, and sensors because the supply chain doesn't exist yet. Tesla is doing the same with Optimus. Does the first company to 100,000 units own the entire stack, or does a component ecosystem eventually emerge?
What happens when autonomous trucks have no cabs? Einride's approach creates interesting questions. Where does the safety driver sit during testing? How do you handle roadside inspections? Does removing the cab make regulation easier or harder?
Join over 4 million Americans who start their day with 1440 – your daily digest for unbiased, fact-centric news. From politics to sports, we cover it all by analyzing over 100 sources. Our concise, 5-minute read lands in your inbox each morning at no cost. Experience news without the noise; let 1440 help you make up your own mind. Sign up now and invite your friends and family to be part of the informed.
Forward this to someone who still thinks flying drones will deliver packages.