How China Caught Up With the U.S. in Drone Warfare | WSJ U.S. vs. China

Military Unmanned Aerial Vehicles (UAVs)
9 followers
1 Video View·Dec 13, 2025
These are videos released by China's state broadcaster.
And experts say it shows the country has made huge strides in developing military
drones for the air, land and sea.
But what's really grabbing experts' attention is this.
Large swarms of systems that show China appears to be equipping these drones
with artificial intelligence.
While we were focused elsewhere in the world,
China saw what they call that window of strategic opportunity
to actually build a force that could compete with the U.S.
China has spent decades modernizing its military with the express goal
of catching up to the U.S.
AI-enabled drones are key to achieving that goal.
While machines that fight entire wars on their own might be a ways off,
experts say limited versions of these drones could be used in battle in just a few years.
So who will make these drones first?
And how will they be used?
The first step of building an AI drone fleet
is to build machines that could run the AI software.
The U.S. has long led the world in advanced military weapons,
with more combat aircraft, armored fighting vehicles
and aircraft carriers than any other country.
To compete, China has copied some of the U.S.'s most powerful systems.
For instance, its Wing Long is comparable to the U.S.'s Reaper drone,
which is used for offensive strikes.
Its Soaring Dragon drone is similar to the U.S.'s Global Hawk surveillance drone.
Like the U.S., China's been building drones that operate across domains of battle.
Both are making unmanned surface vehicles and submarines,
which could be used for surveillance, laying mines or firing torpedoes at warships.
And on the ground, both are developing drones like these robotic dogs,
which could be used to assess threats, fire small arms
and potentially even handle explosives while humans stay at a safe distance.
And then there are the small, cheap systems.
While experts say America leads in large military drones,
China is the world leader in consumer-grade drones,
and it dominates the global supply chain to make these systems inexpensively.
China's DJI controls nearly three-quarters of the global market.
The U.S., meanwhile, has struggled to make reliable small drones at scale,
which is why both sides in the Ukraine war have largely stuck to DJI drones,
though the company says it opposes use of its drones for combat.
Small and mid-sized drones can't fly as far or as fast as large drones,
and they can't carry as much firepower.
But in the future, similar drones could be very useful when linked to other drones.
The U.S. and China are exploring a few strategies for using AI drones,
but they appear to have advantages in different areas.
One of the most sought-after approaches is swarms,
using large numbers of drones in land, air or sea,
all linked with AI, to overwhelm an opponent or carry out a mission.
Even if each individual device has limited capabilities,
swarms work best when some systems can be taken out without compromising the overall mission.
The comparative advantage that China has in the smaller systems,
they could use for things like swarms,
as long as they're able to get the actual technology know-how behind enabling those.
For instance, in this video,
multiple Chinese drones are purportedly working together to locate and take out a target.
Both countries have used drone swarms, but only in training exercises.
In the future, drone swarms may not require human involvement,
but they might be programmed to keep humans in the loop.
U.S. Pentagon policy says humans are responsible for use of force by autonomous systems.
But many worry America's opponents would not follow the same ethical guidelines,
which could enable them to strike faster.
In the United States, we're very concerned about lethal autonomous weapons,
this idea that a drone would operate completely independent from a human
and actually engage targets.
But the problem with all of this is that our adversaries are unlikely
to participate in any kind of ban on lethal autonomous weapons.
And the United States, in fact, is not participating in any kind of ban.
A simpler approach to swarms that's actually been battle-tested is called drone stacks.
Here, different individually controlled drones are brought into a battle space
to perform distinct and often sequential functions,
like target identification from a distance and then confirmation at closer range.
That information can be sent to a human operator who takes out the target.
While AI can be used in limited ways within a drone stack,
like for collision avoidance,
the overall approach is much slower and less sophisticated than a swarm.
And it requires about as many human operators as drones.
Because drone swarms and stacks are largely useful for reconnaissance,
the more sensors they have on board, the more powerful they can be.
Particularly in a hypothetical conflict over Taiwan,
both countries would need as much intelligence about the battle space as possible.
I think this is a key place where they can contribute
in terms of building up our combat capability,
which in turn can bolster our deterrence
and ensure this kind of conflict never happens in the first place.
So I do think that a key capability drones will need to have is sensors
for identifying targets in this highly contested battle space
where we don't expect to have very good ability to see things so we can shoot them.
Those could include standard optical and infrared cameras,
but also things like synthetic aperture radar sensors
that could identify ships and other targets through clouds,
or electro-optical sensors that can see moving targets.
A third approach to AI drones of particular interest to the U.S.
is collaborative combat aircraft,
using AI-equipped drones to support human fighters,
for instance, gathering intelligence or delivering weapons.
That way, the crewed fighter jet would have assistance
and be less of a target than if it were carrying the weapons itself.
Both countries are likely to need different types of AI-equipped drones.
In a confrontation over Taiwan, China could launch drones from its shores,
while the U.S. would need longer-range drones
to deploy from some of its closest military bases.
But both are exploring a huge range of drones
as they try to achieve or maintain a military advantage.
Over the past few decades,
China has been executing a three-part plan to catch up to the U.S. military.
Mechanization, or reducing how many troops it needs
by emphasizing weapon systems.
Informationization, building its surveillance systems
to control battle spaces and link military platforms.
And intelligentization,
speeding up decision-making and communication with AI.
So far, it says it's achieved the first objective.
Truly intelligent drones might take years, if not decades.
I don't think we're going to see AI-enabled drones
in the sense of drones that operate completely independent from humans
in the next five years or so.
But I do think we'll start to see a new generation of drones
that have different characteristics than the Predator and Reaper
that we've all become so familiar with
from the last 20 years in Iraq and Afghanistan.
Eventually, AI drones could replace troops in many contexts.
A U.S. Air Force test earlier this year
put algorithms head-to-head with human fighter pilots in dogfights.
And these systems could massively scale up fleets.
A fighter jet can cost about $100 million.
A DJI drone can cost a few hundred dollars.
But the U.S. has a long way to go
to get these kinds of weapons on the battlefield.
Two key American efforts to develop AI
are the Air Force's Collaborative Combat Aircraft Program
and the Department of Defense's Replicator Program.
But both are just ramping up.
The Air Force is funding two companies
to develop autonomous fighter drones
to fly in partnership with manned aircraft
or connected to commanders on the ground.
But it doesn't expect to receive fully operational drones
until later in the decade.
In the short term,
AI and drones will likely be used in limited ways,
such as more drones being built with autonomous navigation.
When new technology pops up,
the first country that can use it
and then harness that power
generally has a slight advantage,
if not a great advantage.
So timelines are timelines,
but capabilities and actual practical use
of those capabilities is more important.
