The Supercomputer Behind the Stick – Joby Aviation’s Radical Approach to Flight Control

The stick in the cockpit is quietly becoming optional at this point in aviation history, which is happening mostly without much fanfare. In a very realistic, military-tested, real-miles manner, rather than ceremoniously or as a symbolic step toward a driverless future. Compared to most, Joby Aviation appears to have a better understanding of this.

A Cessna 208 flew over 7,000 miles over the Pacific Ocean and the skies above Hawaii in September of last year. The pilot was not there to fly, but to observe. Joby’s Superpilot autonomous flight system, which was controlled from ground control stations occasionally located over 3,000 miles away at Andersen Air Force Base in Guam, handled the actual flying, including the altitude corrections, navigational choices, and moment-to-moment adjustments that prevent an aircraft from becoming a very expensive problem. It recorded more than 40 flight hours. It was flawless. Although that exercise, called REFORPAC, was officially a Department-Level U.S. Air Force drill, what it showed seemed far more significant than any one military goal.

The type of system that doesn’t receive enough attention from those who write about air taxis is Joby’s flight computer, which, for context, operates as three independent units concurrently on their aircraft in a triple-redundant configuration. Range, noise levels, and battery chemistry are frequently discussed topics.

However, the real point of contention is the flight computer, which is what distinguishes a flying prototype from something that could eventually safely transport a stranger across a city. Joby’s S4 aircraft is equipped with four battery packs, six electric motors, and six propellers. Managing all of that in real time requires processing power that wouldn’t be out of place in an enterprise data center, particularly during the transition from vertical hover to forward cruise, a notoriously difficult phase that has humbled more than a few eVTOL programs.

When the NVIDIA partnership was revealed in late October 2025, the market had a better idea of Joby’s true autonomous ambitions. When Joby’s ten years of flight-testing data and FAA certification work are combined with NVIDIA’s raw compute, the result is usually something truly novel. Although it’s possible that the final product won’t resemble either company’s initial vision—that’s typically how these partnerships go—the direction is clear. A system that has already demonstrated its ability to fly itself across an ocean is gradually given more intelligence, computing power, and autonomy.

When you look at the cockpit design philosophy, you’ll notice how purposefully Joby took inspiration from the world of fighter jets. The S4’s sidestick controller manages altitude, attitude, and direction from a single point of contact; it was partially inspired by the F-35B. It’s a decision that shows a sincere conviction that more straightforward, user-friendly controls lower the possibility of human error and, possibly, make the shift to complete autonomy more conceptually clear. There’s a feeling that Joby is creating a car that doesn’t require a pilot, but he is aware that it must persuade the FAA, the passengers, and the Pentagon that, for the time being, having one is still the right decision.

More attention should be paid to the defense angle than is usually the case. In its FY26 budget request, the Department of Defense has allocated $9.4 billion specifically for the advancement of autonomous and hybrid aircraft. That’s not funding for research. That’s money for operational readiness. Joby is in a position that pure commercial eVTOL companies just aren’t, having just finished a successful live-exercise demonstration with AFWERX and Pacific Air Forces. The dual-use approach, which involves developing technology for both the military and the morning commute, is either extremely clever or extremely difficult, and it may involve both at the same time.

The analysts tell a different story than the stock does. For a company that hasn’t yet turned a profit and, according to most predictions, won’t for at least the next few years, Joby’s stock has returned about 172 percent over the last year. There’s a lot of tension there. Investors appear to think that the technology is currently worthwhile. With a consensus price target that is significantly below where the shares have been trading, analysts are a little more measured. It’s still unclear if the market declines or if the reality of commercial air taxis catches up more quickly than the models predict.

A successful Pacific flight, a handshake with NVIDIA, and a cautious speech by a U.S. Air Force officer at a press release about “affordable technologies that support the needs of our Airmen” are just a few examples of the quieter announcements that seem more significant as this develops. The tale of the air taxi has always been loudly promoted. The supercomputer behind the stick, though? Something more akin to silence is being built in that one.