Brandon Tseng (曾國光), the president and co-founder of San Diego-based AI defense startup Shield AI, cuts an unlikely figure for a former Navy SEAL — warm and approachable, nothing like the powerfully built, fierce combatant one might expect of a career soldier, let alone someone who has faced life and death as a special operations warrior.
After graduating from the U.S. Naval Academy, Tseng served in the Navy for more than seven years — first as an engineering officer aboard surface vessels, before being selected for the globally renowned Navy SEALs. He completed multiple deployments to Afghanistan and several locations across the Pacific.
Related report:
Grandfather Was a Nationalist Diplomat — and Called Zeng Guofan "Great-Grandfather"
Tseng is a third-generation Chinese-American. His grandfather fought against Japan in the Second Sino-Japanese War before joining Chiang Kai-shek's Nationalist government and relocating to Taiwan. He later served as a Republic of China diplomat across Latin America for roughly 25 years — postings that took him through Chile, Nicaragua, Guatemala, Venezuela, and El Salvador. He eventually retired in South America before settling in Texas, where Tseng's aunt had established herself, and moved in with her.
That grandfather, Tseng said during an interview at the Taipei Aerospace and Defense Exhibition in September 2025 — where Shield AI made its debut appearance in Taiwan — addressed Zeng Guofan (曾國藩) as his great-grandfather. That makes Brandon Tseng a fifth-generation descendant of the Qing statesman, widely regarded as one of the four most powerful officials of the Qing dynasty's final decades.
Shield AI participated in the Taipei show with its own exhibition booth. Both Tseng and the company's chief technology officer, Nathan Michael, delivered keynote addresses during the event.
"My Grandfather and Father Always Told Me How Lucky I Was to Be Born in America"
Tseng traces his desire to serve in the military to two sources: Hollywood films he watched around the age of 10, and lessons absorbed from his elders about what it meant to grow up in the United States.
Films such asThe Rock andUnder Siege, both featuring Navy SEALs, made a deep impression on him as a child and planted the idea of serving with the country's most elite forces. His family's history reinforced that impulse.
"My grandfather, my father, and my aunt always told me how lucky I was to be born and raised in America — like winning the lottery," he said. "Because they had seen so much war and what happens when the Communist Party takes power. So I developed a sense of patriotism, a desire to give back to the United States and to work alongside the best people. And, of course, the action movies had a big influence on me too."
SEAL Training: 20 Months, 85% Dropout Rate
Tseng graduated from the U.S. Naval Academy with a degree in mechanical engineering and naturally took up the role of engineering officer aboard surface vessels. Yet the dream of joining special operations — planted by those Hollywood films in childhood — remained deeply rooted in his mind, and he eventually volunteered for SEAL selection.
"Once the SEALs accept your application, you go through an extremely grueling and demanding training pipeline," he said. "Qualifying as a SEAL takes about 20 months — nearly two years — and the attrition rate in the class runs as high as 85 percent. That means up to 85 percent of your classmates voluntarily quit. Getting through it is genuinely difficult."
After qualifying, Tseng served in the SEALs for several years and completed multiple deployments to Afghanistan and locations across the Pacific.
Named to TIME's 2025 List of the 100 Most Influential People in AI
TIME magazine included Tseng in its 2025 ranking of the 100 most influential figures in artificial intelligence, published in August. The citation described how a 2015 deployment to Afghanistan crystallized his thinking about the role of AI in combat.
Operating in an urban environment, his team needed to move between buildings with no way of knowing whether enemy combatants were waiting in the next corridor. Tseng concluded that a robot could solve the problem without putting lives at risk. When his overseas tour ended, he decided to build one himself.
The result was Shield AI, which he co-founded in 2015 with his brother Ryan and engineer Andrew Reiter. The company's first product, Nova, was a small autonomous drone powered by AI, designed to navigate confined spaces and detect occupants inside buildings. Shield AI subsequently expanded that technology into a broader family of AI pilot platforms, including one capable of flying an F-16 fighter jet. The U.S. military has since awarded the company multiple contracts.
Founded in 2015: Spotting the Opportunity the Industry Giants Overlooked
Tseng acknowledged that launching a defense startup as a former SEAL was unusual, and aiming to compete in an industry dominated by Lockheed Martin, Raytheon, and Boeing made it rarer still. As a late-arriving small player, where did Tseng find the confidence to take on the industry's giants? His argument, he said, was straightforward: the incumbents were hardware companies, and there was room for a firm that put software first.
"Of course the hardware still has to be built," he said, "but software always has to come first. When I started the company, I believed Shield AI could succeed as a business focused on AI and autonomy — areas those big players weren't prioritizing."
He conceded that the established primes command far greater resources, but noted that the problem is they move slowly and are burdened by organizational bureaucracy — what a smaller company can build quickly is not something a large company can necessarily mobilize to match. Drawing on a comparison familiar to Taiwan's technology industry, he pointed to the Silicon Valley pattern in which nimble startups eventually overtake entrenched leaders.
"That's why you see small tech companies grow into giants — Apple, NVIDIA, AMD," he said. "They were smaller, more agile, able to build better products, partly because they didn't have the same resources as their bigger rivals, and partly because those rivals didn't have the same conviction in their product strategy that someone like Jensen Huang has."
Shield AI Opens Taiwan Office, Strikes Deals with AIDC and NCSIST
Since its founding, Shield AI has expanded well beyond its original U.S. military customer base. The company now serves not only all branches of the American armed forces and the Coast Guard, but has also entered markets in South Korea, Ukraine, and Taiwan. It formally established a Taiwan subsidiary in September last year and has since signed cooperation agreements with two of the island's leading defense organizations.
The agreement with Aerospace Industrial Development Corporation (AIDC, 漢翔) covers building a local support and maintenance ecosystem for the V-BAT, Shield AI's flagship vertical takeoff and landing reconnaissance drone, Tseng said. AIDC is also a potential long-term partner for the X-BAT, which the company describes as the first unmanned jet combat aircraft operated autonomously by AI.
The agreement with the National Chung-Shan Institute of Science and Technology (NCSIST, 中科院) involves licensing Shield AI's Hivemind AI platform, giving Taiwan's drone developers and the Ministry of National Defense access to the same development tools and infrastructure to build AI autonomy systems for their own military unmanned vehicles.
"If you believe that building deterrence is how you strengthen defense, and you want to invest in that — where in the world matters more than Taiwan?" Tseng said.
More in-depth reporting from Storm Media：
- Exclusive | Taiwan's Defense Deadlock and the 'Irresponsible Ally' Warning
- Exclusive | Between Bubble Tea and Missiles: What the KMT-Beijing Meeting Cost Taiwan in Washington
- From New York to Taipei: How Climate Change and Urban Density Are Fueling a Global Rat Surge （Related： Opinion | Eyes on the Horizon: Will Paraguay's Visit Trigger Another PLA Drill? ｜ Latest ）
