Autonomous crop-sprayers lead race to pilot-free commercial aviation

Over an alfalfa field in California's San Joaquin Valley, a small aircraft descends toward the crop with no one in the cockpit. The plane is Pyka's autonomous crop-sprayer, and it is flying lower than any human pilot would dare. Russ Marotzke, a flight test engineer at Pyka, watches from the ground. Flying lower means the spray drifts less, which means less chemical is needed overall—a practical advantage that autonomous systems can deliver precisely because they do not have to account for human comfort or caution.

Pyka operates from a converted World War II hangar overlooking San Francisco Bay. The company builds self-flying aircraft without cockpits, designed to spray crops or deliver cargo. It is one of several firms in a quieter but accelerating race to bring autonomous fixed-wing planes into commercial service. While electric vertical takeoff aircraft—the urban air taxis that dominate headlines—have captured public imagination, the real momentum is building in agricultural and cargo work. Michael Norcia, Pyka's co-founder and CEO, speaks openly about the long-term vision: a fleet of minibus-capacity planes ferrying passengers up and down the US coasts. He believes Pyka may reach that goal before the eVTOL industry does.

About 80 kilometers east of the factory, at a test site reached by a dirt road, Marotzke and a colleague are testing a software update. A dozen Pyka aircraft are already operating in Brazil, spraying cotton and soybeans—work that human pilots once performed. The plane itself is fully electric, with a battery in the nose and a 300-liter tank in its middle section. Its 11.5-meter wingspan makes calling it a large drone seem like an understatement. Inside a shipping container beside the field, engineers highlight on a computer the area to be sprayed. The software calculates the route, accounting for obstacles like nearby power lines that have been mapped into the system. The takeoff is smooth. Fifteen minutes later, after the aircraft senses its spray tank running low, it lands itself, receives a manual refill and a battery swap, then resumes spraying from exactly where it stopped.

Autonomy is not the same as autopilot. Autopilot assists a human operator, much like cruise control in a car. Autonomous systems aim to handle the entire flight—takeoff, navigation, landing—with minimal human intervention, using algorithms to process sensor data and control the aircraft. Self-flying planes have emerged more slowly than self-driving cars, despite operating in what should be a more predictable environment. Mykel Kochenderfer, an expert in safe aviation autonomy at Stanford University, attributes this partly to the fact that major tech companies poured vast resources into cars. But the real barrier is regulatory: aircraft face stricter safety standards than cars, and the consequences of failure are far more severe. Military contracts have helped accelerate the technology, offering fewer regulatory hurdles and sometimes direct customers.

In the United States, Pyka's crop-sprayer is the largest autonomous fixed-wing aircraft approved for commercial civilian use. The authorization came last year, though operations remain tightly constrained to agricultural settings and require a ground operator and visual observer. Brazil approved similar operations under more permissive rules. Pyka plans to scale production from roughly two dozen planes annually to 1,000 by 2030, with each aircraft selling for $550,000. The UK has not yet approved such operations, though the British firm Windracers is seeking permission to launch an autonomous cargo service in Shetland and Orkney. Windracers aircraft are already flying missions in Ukraine, designed to carry goods to remote areas.

Companies pursuing autonomy are taking different paths. Pyka and Windracers build aircraft from scratch, arguing this allows autonomy to be designed into the system from the beginning. Others retrofit existing certified planes. Reliable Robotics, backed by Boeing's investment arm, is testing its system on the Cessna 208B Grand Caravan, a single-pilot cargo plane that can carry about 1,360 kilograms over hundreds of kilometers. Retrofitting lets the company focus solely on proving the autonomous system's safety rather than seeking approval for an entirely new aircraft. Merlin Labs is working its way up through progressively larger military aircraft and is now applying its system to the two-pilot Lockheed Martin C-130J military transport plane, with commercial multi-crew cargo planes as the next target.

The companies diverge sharply on artificial intelligence. Reliable Robotics avoids AI altogether, arguing it would complicate certification. Merlin takes a far more AI-centric approach. This divide appears most clearly in detect-and-avoid systems, one of autonomous flight's greatest challenges. A pilot can spot and maneuver around other aircraft and obstacles; an autonomous system must replicate this with virtually no margin for error. Reliable has added forward-looking air-to-air radar to detect other aircraft more than eight kilometers ahead, with software that follows fixed rules for response. Merlin uses AI-powered cameras to detect and classify objects. Pyka has used lidar from the start to detect trees, vehicles, large birds, and terrain, but lidar has limited range, so Pyka plans to add AI-powered cameras for the first time. As Norcia explains, there is no need for AI in most functions, but distinguishing six pixels in the distance that represent an airplane from other visual noise is precisely where AI excels.

The question of how autonomous aircraft will communicate with air traffic control in shared airspace remains unsolved. Reliable's approach is to station a remote pilot on the ground, initially fully trained, to handle radio communications and safety-critical decisions. Merlin plans to use generative AI trained on thousands of hours of recorded air traffic control exchanges to interpret instructions and respond autonomously. Merlin's problem is harder, George acknowledges, but the company wants to move beyond remote piloting entirely. Merlin intends to reduce pilots in stages, from two to one to eventually none. Pyka, for now, is content to let others blaze the trail in shared airspace operations. Pilots' groups remain skeptical. The US Air Line Pilots Association calls removing pilots a serious gamble with safety. The National Agricultural Aviation Association, representing crop-dusting pilots, notes that small uncrewed aircraft are hard to see and that piloted planes can spray larger areas faster. Yet even skeptics acknowledge that if autonomous systems do make their way into commercial aviation, they could make piloted flying safer by introducing more automated safeguards into the broader system.