How it must have stung Lockheed Martin this spring when it was announced that the first of the next-generation stealth fighters for the U.S. Air Force would be built not by them—the company that practically invented stealth—but by Boeing, a company recently known more for its missteps than successes.
After all, wasn’t it Lockheed—with its legendary Skunk Works—that created the first almost radar-invisible plane, the F-117 Nighthawk? Didn’t Lockheed also develop the F-22 Raptor and F-35 Lightning II? Lockheed’s dominance in stealth aircraft seemed unquestionable. Yet, there stood President Trump, announcing that the Next Generation Air Dominance (NGAD) fighter would be called the F-47 (with Trump serving as both the 45th and 47th president) and built by Boeing, leaving defense insiders, analysts, and Lockheed’s own engineers completely befuddled.
The Digital Fighter
But this may not have been about building aircraft, which one might argue Boeing has done a poor job of lately; it may have been about how you design one.
Though there was no official reason given, according to Air & Space Forces magazine, a report by SimpleFlying speculates the real reason for Boeing’s surprise victory was its “digital-first” approach and maintains that the U.S.Air Force wanted not another prototype to be hammered into shape after years of wind-tunnel trials and flight tests, but a fully realized digital twin—a fighter that could exist and evolve virtually long before the first metal part is shaped.
Lockheed issued no formal protest after the contract was awarded to Boeing. Nevertheless, the company must have been thoroughly disappointed from top to bottom, especially when it has gone to great lengths to prove it is also digitally adept. Lockheed has given presentations at CAD conferences and issued press releases about designing the B-21 stealth bomber from scratch using the latest digital tools. Siemens Digital Industries Software announced in early 2002 that Lockheed Martin was going to use its Xcelerator portfolio for “current programs and new initiatives.”
Kelly’s Ghost
Lockheed, with its impressive stealth pedigree and a decent digital toolbox, thanks to Siemens, may still need to shake off the old-school reputation it acquired during its Skunk Works days. Could it have been the ghost of Clarence “Kelly” Johnson, the legendary leader of Lockheed’s Skunk Works, that cost Lockheed the F-47 contract? The U-2, the SR-71, and the first stealth fighters, the F117, were all innovations, but they had been the product of Kelly’s sheer will and determination, rather than of software and data.
Boeing, by contrast, could have pitched itself as a simulation-driven, model-based engineering program, digital twin operation for as long as the technology has existed. Boeing’s digital heritage dates back to the 1990s when the 777 became the world’s first airliner designed fully in 3D CAD using Dassault’s CATIA software. At the time, Boeing’s engineers called it “paperless design,” taking pride in having every bracket and bolt first exist in virtual space.
Kelly’s ghost was not in the room, according to Space & Air, which states that past performance accounted for “less than 10 percent of the scoring,” though that was more likely to be a reference to Boeing’s “issues” and Lockheed’s cost overruns and delays on the F-35 program rather than Kelly’s ghost.
Turnaround for Boeing
Boeing’s turnaround in fortune after winning the F-47 contract is nothing short of head-snapping. The deal is worth a total of $47 billion over the next five years.
Boeing’s 737 MAX fiasco may have been the nadir of the fabled company’s recent tortured past. Two crashes—in Indonesia in 2018 and Ethiopia in 2019—killed 346 people and grounded the global 737 MAX fleet for nearly two years. Investigations revealed that Boeing had concealed critical information about a new flight-control system, MCAS, which could cause the nose to pitch down based on faulty sensor data. The company was accused of prioritizing speed-to-market and cost-cutting over safety and transparency, as chronicled in the 2022 Netflix documentary Downfall, the Case Against Boeing. An FAA investigation led to criminal charges, billions in settlements and a reputational collapse. The MAX crisis exposed deep cultural and engineering flaws within Boeing, eroding trust with regulators, airlines, the flying public, and investors, most of whom had effectively written Boeing off. Boeing was to fall to 2nd place in commercial aviation, behind Airbus.
As if losses in Boeing’s commercial aircraft division last year were not enough, Boeing’s Defense, Space & Security division lost even more: roughly US $6 billion. This was after the CST‑100 Starliner encountered enough technical setbacks for NASA to give the job of ferrying astronauts to and from the International Space Station to SpaceX.
The Rise of the Algorithmic Designer
Living in the world of design and engineering software, as we are, things have indeed gotten more exciting. With new applications, new computing paradigms and AI, the acceleration from experiment to simulation has increased and design exploration itself has taken on a whole new meaning.
Among those leading the charge is Bradley Rothenberg, CEO and founder of nTop—and a self-professed aircraft-design obsessive. Brad, as he likes to be called, has positioned nTop as a tool capable of transforming how aircraft shapes are designed. He boldly asserts that nTop’s computational design—the ability to automatically generate and test thousands of configurations—is exactly what’s needed to optimize aircraft design.
Brad’s argument centers on engineers who have created one shape that works, that is altogether familiar to everyone who has seen a commercial passenger jet, and are content to use that shape, with the slightest of modifications, from now on. “It works, so why fix it?”
Yet the military, whose operational envelope and missions change often, must adapt aircraft shapes accordingly.
For military applications, nTop’s algorithms, which can spin out a thousand shape variants in minutes, make perfect sense. nTop can optimize simultaneously for multiple criteria: aerodynamic performance, radar cross-section, structural weight, thermal loading, and even manufacturability. This isn’t CAD in the traditional sense, that which takes a concept already formatted in an engineer’s head, and gives it precise shape—it’s real conceptual design by computation, driven by constraints and objectives.
Brad has hosted workshops where engineers are challenged to find the best possible wing or fuselage using nTop’s platform.
When you consider the complexity of a stealth aircraft—its surfaces engineered not only for lift but also to deflect electromagnetic radiation—it becomes clear why digital exploration has an advantage. Each surface angle, panel joint, or antenna placement can affect radar signature, heat dissipation, or signal interference. Manually adjusting and testing each variable would take years. Computational design can do it while you make yourself a coffee.
Today’s stealth aircraft must be designed not by intuition but by iteration. There are simply too many possibilities to consider and simply finding one that works is no longer good enough. Digital twins simulate the behavior of the aircraft in every conceivable flight regime and threat environment. AI-assisted analysis compares thousands of variants for aerodynamic and electromagnetic efficiency. Furthermore, AI will be called on for what the Air Force calls “manned-unmanned teaming,” AKA one manned fighter jet quarterbacking a team of drones.
Digital Design as Corporate Redemption
No one could argue that Boeing desperately needed a win. It may be more likely that the Air Force picked Boeing not so much on technical merit but to keep it alive. Military aircraft manufacturers are few. One less aircraft manufacturer would have put the U.S. weapon procurement system, which relies on competition between manufacturers, out of whack.
Lockheed-Martin should take heart. It is, after all, the company that gave us the Skunk Works and the SR-71, which, almost 50 years later, holds speed and altitude records (2,193 mph and 85,069 ft) for crewed air-breathing aircraft.
Lockheed is not out of the picture. The system will come back to them, predicts a former procurement chief, noting that Boeing has only received the first “increment” of the award.
Here it Comes: AI Role in Design for the Defense Industry
The message for the entire defense industry is clear: in order to compete, you must present yourself as a company that designs digitally. Whether you’re building jet fighters, EV batteries, or satellite constellations, the pattern is the same—fast, cheap, digital iteration beats slow, expensive, physical experimentation. Digital exploration enables broader trade studies, more optimization, and fewer surprises on the production floor.
Now that all major military aircraft manufacturers—including Boeing, Lockheed Martin, and Northrop Grumman—have adopted similar state-of-the-art design, simulation, and manufacturing platforms, any true technological edge is increasingly difficult to establish. With each company boasting its own digital twins, debates over whose tools are superior become more subjective than substantive. To the agency officials tasked with determining who has the technology edge, all the major players appear to hold similar cards.
Obviously, radical new technology, such as AI, could completely shuffle the deck —if it works. This could lead to not a distant future in which the defense contractor that has the most AI tools holds the most aces.
And so, the race is on. All must develop or acquire applications such as nTop, which is perfectly capable of simulating a thousand versions of a wing, or even a whole aircraft (exterior only, at a low level of detail), and rank them in order, judging from one or more criteria. That nTop does this without AI as it is being used today {nTop uses optimization routines, not neural networks) makes us wonder what heights it could achieve with an AI infusion.
AI implementation and adoption could lead to as radical a change as the one that took place with stealth aircraft. The first stealth aircraft, the F-117 Nighthawk, took its shape, which is very angular, composed almost entirely of flat, triangular panels, from the computational limits of the 1970s, when it was designed. At that time, computers couldn’t handle the complex calculations needed to predict how radar waves would scatter off curved surfaces. So, Lockheed engineer Denys Overholser, working under Kelly Johnson, utilized a Soviet mathematical paper by Pyotr Ufimtsev that demonstrated how radar reflections could be modeled on flat, planar surfaces. Overholser and his team wrote code—called Echo 1—that could compute radar returns only from flat triangles.
Later stealth aircraft, such as the B-2 Spirit and F-22 Raptor, adopted a smooth and curved shape, a testament to the additional computing power available in the 1990s, when computers were powerful enough to model radar reflections from complex, curved surfaces.
What we see now of the shape of the F-47 might as well be called, at best, extensions of present designs or, at worst, artist or AI hallucinations. The eventual shape found after millions of iterations could very well be one that no human has devised.