Drones have become Ukraine's weapon of choice against Russia, and have helped Iran withstand seven months of heavy US pressure. Facing its own David versus Goliath odds across the Taiwan Strait, Taiwan is now looking at drones as a new hope for closing the gap in military power with China. The US military has floated a "Hellscape" strategy for the strait, but retired Admiral Lee Hsi-min, who proposed Taiwan's own "Overall Defense Concept" back in 2017, has remained a sober and consistent skeptic of the rush to treat drones as a cure all for Taiwan's defense.
Lee commanded the submarine Hai Hu on the front lines of the 1996 Taiwan Strait crisis and later became the architect of the Overall Defense Concept, Taiwan's asymmetric approach to denial based deterrence. In an interview with Storm Media, he laid out, in detail, where he believes Taiwan's defense revolution needs to go. Drawing on decades of practical military experience, he picked apart the US Indo-Pacific Command's "Hellscape" concept and argued bluntly that "the best way to defend Taiwan has to be rooted in Taiwan itself, not in America."
On the government's push to build up a drone force, Lee said the defense ministry has gotten the order of operations backwards, lacking any overarching operational doctrine. Without first clarifying the warfighting concept and the battlefield mission a drone is meant to serve, he warned, jumping straight to spending money on hardware only turns procurement into a public relations exercise for arms dealers and a race to the bottom on price. More dangerous still, he said, is that Taiwan has a "blank brain" when it comes to military AI, since its defense research still relies on American command and autonomy software, leaving Taiwan exposed to a "sovereign AI" crisis if a foreign supplier ever cuts off access.
Hellscape Is Just His Own Overall Defense Concept By Another Name
Lee said future warfare will be dominated by the combination of unmanned systems, artificial intelligence and space based networks, and that once the three are tightly integrated, "there's no reason to keep investing heavily in traditional manned platforms." He pointed to the Russia Ukraine war as proof that pairing drones with AI is workable, but noted that real combat also exposes bandwidth, low latency transmission and jam resistant communication problems, which is exactly where space based assets become critical. Whoever masters all three, he said, is the one who can actually realize a "Hellscape" on the battlefield.
Asked about the "Hellscape" concept for the Taiwan Strait put forward by Admiral Samuel Paparo, head of US Indo-Pacific Command, Lee argued that Hellscape is, in essence, exactly what he has long advocated as the Overall Defense Concept, or ODC: in the simplest terms, using "a large number of small, mobile, dispersed, precise and lethal things" to achieve denial based deterrence.
Why The US Military Cannot Actually Pull Off Hellscape
Lee said flatly that nobody has pulled off Hellscape yet, and while it may eventually be achievable with continued effort, having the US military deploy it across the Taiwan Strait is "pretty much not going to happen." The reason is not a lack of American technology, he said, but a lack of bases and sustained logistics. Small unmanned platforms have limited range and cannot operate far from resupply points, and even with US bases in Japan or the Philippines, those are still too far from the strait. "These small platforms can't rotate on patrol the way a warship or fighter jet can, and they can't loiter for long. If the Chinese military hasn't launched a full scale invasion yet, how do you pre position a 'Hellscape' in the strait in advance? And how would you even sustain it?"
Lee said the entity best positioned to actually realize a Hellscape in the Taiwan Strait has always been Taiwan itself. When he discusses the issue with foreign friends who follow Taiwan's situation, he said, he always emphasizes that "the best way to defend Taiwan is from Taiwan, not from America." Taiwan has the advantage of proximity and a range of communication options, and does not have to rely on fragile space based or long range relay links. China, conversely, is geographically far better positioned than the United States to deploy a Hellscape of its own across the strait, precisely because it lacks the forward basing problem Washington faces.
As for the US military's own planning, Lee believes Paparo's initial remarks on Hellscape were "overstated," and that subsequent comments from him have grown visibly more cautious after deeper staff level wargaming, "for the simple reason that the US military itself cannot actually do it." He also noted that the Pentagon's Defense Innovation Unit once pushed the Replicator Initiative, aimed at rapidly mass producing low cost unmanned systems to offset China's numerical advantage, but that with Donald Trump's return to the White House and a shift in policy priorities, few in Washington still talk about it.
Buying Drones Is Not A Strategy
On a Center for a New American Security report released this year outlining a Taiwan Strait Hellscape extending 80 kilometers offshore in four layers of unmanned defenses, with the goal shifting from merely stalling China for a month until reinforcements arrive to independently defeating a Chinese assault without US intervention, Lee was blunt: "So in the end, to put it bluntly, Taiwan's example shows you simply cannot rely on America. You have to rely on yourself."
On the military's recent buildup, Lee said the push toward a littoral combat command, decentralized command and control, and mission type orders all reflect ODC thinking at their core. Even so, he argued the defense ministry's current approach has fallen into the trap of lacking any overarching operational guidance.
The legislature passed a roughly NT$240 billion unmanned vehicle development package over six years, but Lee had sharp words for it: "You're just throwing money at the problem, and the problem won't solve itself because of that." He pointed to American contractor Anduril Industries' aggressive promotion of its Altius drone system in Taiwan, which produced the spectacle of "arms dealers posing for photos with the defense minister," while there is still no clear roadmap for what kind of autonomous unmanned combat ecosystem Taiwan is actually trying to build.
Asked whether Taiwan's growing drone force is meant to stall for a month while waiting on American help or to independently deny a Chinese assault, Lee did not give a direct answer. "If you don't have a strategic doctrine and you just start buying drones, you'll inevitably lose your sense of direction. What we're missing most is that nobody really knows what Taiwan's drones are actually supposed to be used to fight. Nobody seems to know."
Lee warned more seriously that without operational guidance, jumping straight to procuring finished products under the traditional lowest bid system creates a vicious cycle: winning contractors, chasing profit, quietly mix in low cost Chinese made components to pass acceptance inspection, only to have the goods rejected later by the military. "That cycle never ends. How long is it going to take before Taiwan's drone combat power is actually built up?" Lee said unmanned systems obviously matter, but "buying drones" is not itself a strategy. Taiwan first needs to decide how it intends to fight, and let that operational concept determine organization, software, AI, communications, production capacity and weapons.
A Sovereign AI Crisis Hiding Inside Taiwan's Drone Program
As the public pins hopes on drones becoming Taiwan's "next silicon shield," Lee argued bluntly that "Taiwan's drones are actually quite weak," and that the weakness is not in hardware but in the lack of core components and a combat brain system. He disclosed that the National Chung Shan Institute of Science and Technology still relies on American contractor Anduril's Lattice command and control platform and Shield AI's Hivemind autonomous flight software when training drones. "Taiwan should be developing this on its own," he stressed, retaining control over core software, data and systems, "rather than simply adopting someone else's."
Lee said that as generative AI rapidly advances toward "agentic AI," AI's role in the military is no longer merely assistive, and can even take the lead on mission planning and dynamic command and control, which directly implicates what he calls Taiwan's "sovereign AI" crisis: "If one day the US side simply shuts off system access, does the force you spent a fortune building just collapse entirely?"
He pointed to Ukraine as a cautionary example, noting that Elon Musk previously restricted Ukrainian forces' use of his Starlink satellite network in certain operational areas over geopolitical concerns, hampering a Ukrainian offensive. Lee warned that if Taiwan's combat software, AI models and communications are entirely controlled from outside the country, it would face enormous strategic disruption risk in wartime.
Why The Non Red Supply Chain Debate Misses The Real Risk
A recent report from the Taiwan based Research Center for Digital, Society and Emerging Technology found that Taiwan's drone supply chain remains heavily dependent on China for upstream materials such as battery cells and rare earth magnets, and recommended further expanding Taiwan's "non red" supply chain. Asked about China's advantages in drone manufacturing and swarm coordination, and about the controversy over a non red supply chain for drones, Lee cautioned against letting political ideology hijack the issue.
Lee argued that who manufactures a drone's airframe or basic modules barely matters, "as long as you make sure you have backup alternatives for when wartime supply is cut off, you can even treat the other side's commercial components as something you put to your own use." The real key to defense strategy, he said, lies in the "brain level decision system" and jam resistant communications and control, including autonomous flight control, target identification and mission command software. If access to those is shut off in wartime, every unmanned platform would be wiped out at once. In Lee's view, ordinary commercial grade components, even those sourced from China, pose a manageable risk as long as Taiwan stockpiles in advance and builds alternative sources; but flight control, mission planning, communications and AI models controlled externally for updates, authorization or access would directly determine whether these weapons still function once a war starts.
On industry hopes that Taiwan can become a global defense supply chain hub through non red sourcing, Lee argued the government's biggest strategic blind spot is conflating national defense readiness with industrial policy. "Being part of the global non red supply chain is fine from a commercial and industrial standpoint, but does being part of a supply chain directly defend your country? No, it doesn't." He questioned whether Taiwan, if it wants to fully develop its own weapons while completely excluding red supply chains, can realistically clear the technical and cost bar required. A vibrant private sector needs ten or twenty years of sustained, regular government budget support if it is to be leveraged for defense, he argued, not short term special budgets renewed every few years, and dressing up defense procurement as industrial policy only weakens the substance of national security.
Why Ukraine's Experience Cannot Simply Be Copied Onto The Taiwan Strait
In both the Russia Ukraine war and the recent US Iran conflict, unmanned systems have become a genuine game changer, and there has been intense interest in whether "the Ukraine experience" can be replicated across the Taiwan Strait. Lee quoted Prussian Field Marshal Helmuth von Moltke the Elder's famous line that "no plan survives first contact with the enemy," arguing that no battle plan can simply be copied wholesale onto another conflict.
Some Western scholars have questioned whether Ukraine's experience, fought largely across open plains with a far calmer Black Sea than the Taiwan Strait, can really be transplanted to the strait, where Ukrainian style naval drone raids against Russia's Black Sea fleet would be much harder to pull off. Lee countered that the strait's conditions are also a natural shield for Taiwan's defenders. "I agree that small unmanned boats can't operate in rough seas and high winds. But rough seas and high winds are also a problem for the Chinese military. If they can't get across, then I don't need to send out counter drone boats either."
Lee stressed that the logic of defensive warfare is that Taiwan reacts only once the enemy moves. Any Chinese amphibious landing would need to pick a window of calm sea conditions, which is exactly the tactical opening for Taiwan's small, mobile, dispersed unmanned platforms to swarm out and conduct lethal interdiction. He further noted that the real obstacle to transplanting Ukraine's drone tactics directly is that Taiwan's dense urban high rises and complex city terrain constrain the communications, navigation and operating space many drones need.
Software Not Hardware Is What Makes A Drone Force Fast
Former CIA director and retired general David Petraeus and Atlantic Council fellow Clara Kaluderovic wrote in Foreign Affairs that Taiwan should pair world class engineering talent directly with its military units for brainstorming and collaboration, using intensive exercises to build a rapid iteration system similar to Ukraine's. Lee's view differs somewhat.
"Bringing in engineers alone doesn't help, because engineers think differently than you do about fighting a war," Lee said bluntly, arguing that civilian engineers focus on technical processes and simply do not understand the combat logic a frontline soldier applies on the battlefield. He said he felt this acutely while serving as head of the defense ministry's systems and information division: "We grew up in different environments. We think about how to fight a war. They think about engineering." Lee said the most valuable lesson from the Ukrainian battlefield is the speed of "software defined" iteration: "This week you use a counter drone system to knock my aircraft down, and next week you can't knock it down anymore. Why? Because everything now is software defined, so I can change fast."
Lee criticized Taiwan's defense industry mindset for staying locked into outdated, hardware bound architectures, with a single combat system often going ten or twenty years without an update. "If you rely only on hardware, your iteration slows down, so what you should be doing now is replacing as much as possible with software. Just like nobody swaps out the parts inside their phone anymore, you just download new software to run new things, and drones should work the same way." He said the government's core task is to build a dialogue mechanism linking frontline soldiers, systems architecture and civilian software talent, integrating top to bottom so each side understands the other's thinking. What Lee proposes is, in fact, fairly close to what Petraeus and Kaluderovic suggest about pairing engineers with frontline operators, but places more emphasis on the software layer and on the different ways soldiers and engineers think.
Taiwan's Weapons Programs Lack Integration Not Money
On material readiness for a potential Chinese assault, Lee said energy independence is undoubtedly Taiwan's biggest vulnerability in the event of a blockade, but narrowing the focus to weapons and equipment, Taiwan's most fatal weakness is its lack of independent defense production capacity. Critical raw materials or rare earths can be stockpiled in peacetime, he said, but the real test is whether Taiwan, like Ukraine, can keep producing and replenishing drones and other asymmetric platforms locally once large numbers are destroyed in wartime.
Asked about the staying power of defense production bodies such as the National Chung Shan Institute of Science and Technology, Lee said bluntly that the core problem "isn't money," but that the entire chain from combat concept to production has never been integrated. He pointed to Israel as an example, where defense research and development from day one has paired soldiers with real combat experience alongside engineers; by contrast, he said, the institute's ranks are made up almost entirely of "technical officers" and research PhDs, and after being converted into an administrative legal entity it has even less military field experience, leaving its researchers with little understanding of what frontline units actually need to survive and fight on the battlefield.
Lee recalled that when he was promoted to major general and took charge of the defense ministry's systems and information division, the unit was full of information and systems engineering PhDs, yet the "battlefield" an engineer imagines and the one a frontline commander understands are worlds apart. He also shared a telling anecdote: he once confronted the institute directly, asking why it could produce the Hsiung Feng II and III missiles and the Sky Bow missile but not a Stinger missile. The researchers' answer, he said, was simply, "nobody gave us the requirement," a clear sign of how disconnected operational needs are from the research system.
On the question of indigenous defense and conventional platform procurement, Lee acknowledged that Taiwan's international political reality leaves little market for traditional mainline weapons such as fighter jets or large warships, since "nobody dares to buy them even if you build them, so there's no market abroad, and the domestic market is far too small to sustain the effort." Only asymmetric unmanned systems, he argued, with their dual use civilian and military nature and comparatively flexible technical barriers, have a genuine shot at being self sufficient or even exportable.
Lee also pointed to the small, fast missile assault boats he pushed for during his tenure as Chief of the General Staff, a program that was scrapped after internal military opposition, only for the force to now return to championing unmanned boats after the Russia Ukraine war validated similar concepts. He lamented that the small, dispersed thinking behind that earlier program has much in common with today's push for unmanned vessels, and that the military's reversal and reversal again reflects the cost of strategically outdated thinking, wasting a decade of potential iteration and improvement.
What Actually Decides Asymmetric Wars
On Taiwan's psychological defense and social resilience, Lee acknowledged that will to fight is extremely hard to measure, but said Taiwan's prolonged internal infighting and talk of defeatism pose a serious threat to national security. "If you let the other side believe Taiwan doesn't want to fight at all, they'll conclude that the moment they act, you'll collapse, and that makes them more likely to act. Either way, both sides end up losing."
Lee pointed directly to what he sees as the flaw in current political mobilization efforts, saying the ruling party "has to make the whole public genuinely feel that 'we are fighting for Taiwan,' not 'fighting for the Democratic Progressive Party.'" If resistance gets framed as fighting for a specific party, he warned, that effectively destroys Taiwan's will to resist, and political infighting does nothing to build genuine consensus on defense.
Lee also invoked Carl von Clausewitz, the Prussian military theorist often called the West's sage of war, and his famous line that "war is the continuation of politics." He argued that Taiwan's biggest current danger is having completely shut the door on political negotiation, and that neither the strong nor the weak should ever close off that channel entirely.
In Lee's strategic thinking, stopping an enemy at the beachhead is not the end of a war but a way of preserving Taiwan's leverage at the negotiating table. The US campaign against Iran demonstrated overwhelming precision strike capability, he noted, yet struggled to translate that into a real political endgame; even as Iran and the US clashed fiercely, "underneath the surface, they kept talking, fighting hard while still negotiating quietly behind the scenes." By the same logic, he argued, if Taiwan's military can repel an enemy's first wave at the beach but lacks a political negotiating strategy with real strategic depth, the cross strait standoff will remain unresolved, because "stopping the enemy at the beach is ultimately about earning a seat at the negotiating table."
Asked whether the fact that neither Ukraine nor Iran has been quickly crushed by a militarily superior opponent in today's two defining asymmetric conflicts amounts to genuine strategic victory, Lee did not answer directly, instead returning to basic material and psychological factors. On the material side, he said modern attrition warfare ultimately comes down to resource reserves and staying power: "The reason we want low cost and large quantity is that quantity sometimes becomes quality. Once your numbers reach a certain scale, that scale itself becomes a form of quality in the contest." He noted that American industrial output, not just technology, overwhelmed the Axis powers in World War Two, and that Ukraine's battlefield today is similarly mired in a resource based war of attrition.
On the psychological side, Lee stressed that willpower can sometimes override battlefield logic, pointing to how US forces held the advantage in most engagements during the Vietnam War yet were still forced to withdraw as domestic antiwar sentiment eroded political will at home. Whether a weaker state can hold firm through sheer resolve, and whether a stronger power's political will can endure a long war of attrition, is, in Lee's view, the real question that will decide the outcome of today's asymmetric conflicts.
YP