Aaron
Nathan

Somewhere between a Cornell engineering lab, a DARPA-sponsored autonomous car race through Pasadena, and a critical piece of technology swallowed by Apple, Aaron Nathan developed a very specific opinion about the future: the machines that will run the world need to know exactly where they are. Not approximately. Not within a car-length. Exactly.

That conviction is the entire engine of Point One Navigation, the San Francisco company Nathan co-founded and continues to lead as CEO. The company does one thing with unusual rigor - it tells machines their precise location, to the centimeter, in real time, at scale. The product underneath that promise is Polaris, a professionally managed RTK (Real-Time Kinematic) corrections network that layers over GPS signals the way a precise ruler layers over a rough measuring tape. Where consumer GPS has an accuracy window of a few meters, Polaris narrows it to centimeters. For an autonomous lawnmower, a precision agriculture drone, or a self-driving delivery vehicle, the difference between "a few meters" and "centimeters" is the difference between working and not working.

In November 2025, Khosla Ventures led a $35 million Series C into Point One - a round Nathan described as oversubscribed, meaning investors wanted in faster than the company needed them. Total capital raised now sits north of $52 million. For a 55-person team, those are serious numbers. They reflect a growing conviction in the market: that precision location is infrastructure, not a feature.

Nathan arrived at this problem through one of the more interesting routes in the industry. At Cornell, he didn't just study autonomous systems - he drove one. As staff researcher and team lead at the Cornell Autonomous Systems Lab, he helped build and pilot the team's entries for the 2005 DARPA Grand Challenge and the 2007 DARPA Urban Challenge. That vehicle - a Chevrolet Tahoe the team named "Skynet" - became part of robotics lore when it collided with MIT's autonomous Land Rover "Talos" during the Pasadena race. The incident was later published as an academic case study on autonomous vehicle behavior. Nathan was there for the whole thing.

After Cornell, he moved into precision navigation at a professional level, joining Coherent Navigation as Chief Architect. Coherent Navigation was working on high-precision GNSS technology - the same general territory Nathan would later build Point One around. That company was acquired by Apple, which means Nathan's early work on precision location is baked somewhere into the world's most widely deployed consumer mapping platform.

Between Coherent Navigation and Point One, Nathan made a detour into the enterprise software world - but kept his founding instinct sharp. He co-founded adeptCloud, a secure cloud file-sharing startup, serving as CTO. In 2013, Hightail (formerly YouSendIt) acquired adeptCloud, and Nathan joined as VP of Technology. There, he led the launch of Hightail Spaces, scaled infrastructure, rebuilt desktop software, and oversaw a 100-plus person engineering team. It was useful experience for someone who would later need to build a company that serves both solo developers and large OEM partners.

He co-founded Point One Navigation around 2015 with Bryan Galusha, another Cornell alum who had worked on the DARPA Urban Challenge. They started from a shared frustration: the technology to achieve centimeter-level precision location existed, but accessing it required deep expertise, expensive hardware, and integration work that put it out of reach for most product teams. The mission became making that accuracy accessible - not just technically but operationally.

Point One's architecture reflects Nathan's "Intel Inside" instinct. The company doesn't sell to consumers. It sells to OEMs, developers, and industrial operators who embed Polaris and the Atlas INS positioning engine into their own products - autonomous farm equipment, delivery robots, construction machinery, precision drones, and survey systems. The customer experience is their product. Point One is the invisible enabler underneath it.

The technical edge Nathan talks about most is not the RTK network itself - several companies offer corrections services - but the combination of network and software. Point One developed "online converging calibration" algorithms that automatically learn how a sensor is oriented when installed on a vehicle, eliminating the manual calibration step that has historically made precision navigation deployment slow and expensive. "If you put your IMU on a rate table and calibrate all those biases, that doesn't help you understand how that IMU is oriented to the front of the vehicle," Nathan explained in an interview with Inside Unmanned Systems. The algorithm solves that automatically. It's a small sentence with enormous implications for how quickly a manufacturer can go from "bought the hardware" to "shipping the product."

Nathan's market vision lands in three areas he's consistently excited about: the alignment of the digital and physical worlds, handset-level precision applications, and the broader robotics industry - which he views as remarkably early-stage for what it's becoming. Autonomous lawn equipment, UAS delivery, precision agriculture, construction robotics - each of these sectors needs what Point One sells. The Series C from Khosla is explicitly aimed at accelerating infrastructure expansion, international operations, and deeper OEM integration across all three.

At 55 people and with $52 million behind them, Point One sits at a particular inflection point: past the "is this technology real" phase, now fully in the "how fast can we scale this" phase. Nathan has done the early-stage company thing twice before, and he's been through an acquisition twice. The difference with Point One is that he's not building to be acquired. He's building the company that acquires would-be competitors as customers.

What makes Nathan an unusual CEO for a deep-tech navigation company is that he's not purely a hardware person. His background spans embedded systems, cloud infrastructure, and enterprise software. At Hightail, he was thinking about latency optimization and continuous integration pipelines. At Point One, he built a developer-first API with GraphQL access, real-time device observability, and a subscription correction service - tools that feel more like a software platform than a GPS company. That blend is deliberate. The company's thesis is that the hard part of precision location is not the physics. The hard part is making it programmable, deployable, and reliable at production scale.

Nathan's current challenge is not technical validation. It's category definition. "Physical AI" - the term Point One uses for the layer of autonomous machines operating in the real world - is still coalescing as a concept. Autonomous vehicles, delivery robots, precision agriculture systems, and industrial drones are all arriving in commercial volumes at roughly the same time. They all need what Point One makes. Nathan's job is to be in the room when those procurement decisions happen, and to have the network and the reputation to be the default choice. An oversubscribed Series C from Khosla Ventures is a strong signal that he's getting there.