China Girl, Part 1: China’s gender arbitrage in defense tech - why China fields 10x the amount of drones America does (hint: they hire women)
Examining how China mobilizes women in defense technology to innovate, outpace, and beat the U.S. through gender parity in labor via legal, economic, and production scale
Part 1’s subsections:
The legal situation: China’s dual use technology approach via military-civil fusion vs U.S. constitutional barrier
The numbers situation, or gender arithmetic: comparing China’s labor stats with America’s in defense
What China’s utilization of women produces at scale: drones, AI, and weapons
In this first part of the China Girl series, I begin by looking at the legal and regulatory frameworks that shape both China and the US’s respective labor pipelines for defense technology. I then examine the labor pipelines, their economics, and the measurable strategic production outputs of each.
This four-part series will examine how innovation economics and competition policy intersect with gender as a strategic imperative in defense technology, and why this matters from a forward looking perspective especially in regards to AI development and governance. Part 1 here quantifies China’s structural advantages through broader workforce integration and the measurable innovation outcomes that result. Part 2 will analyze the policies and institutional failures causing America to voluntarily restrict its own talent base through cultural barriers, policy constraints, and bureaucratic inertia. Part 3 will examine the historical precedents of women in defense and the economics involved. Finally, Part 4 extracts lessons for AI governance, where identical vetting requirements and talent access challenges will determine which nations can develop advanced AI systems at the scale and speed required for strategic competition.
Before we get started, some groundwork as I think it is necessary to make clear what exactly is meant here. When I say innovation economics, what is that? What does that even mean? Since I basically have half a PhD and despise jargon, let me just say, excellent question, dear reader. If you didn’t ask it, now you know it is an excellent question to ask.
So, innovation economics, what is it?
Innovation economics is the conceptual framework for how I’ll approach practically analyzing (1) how labor market dynamics are a competitive factor, and thus (2) how the structural advantages of each labor market form the de facto competition policy of the countries. I then use the information from the labor market dynamics and competition policy along with projections/current stats to quantify them and ultimately measure the economic impact of these things.
By quantifying measurable outcomes, it becomes easier to identify some emerging questions to consider, in this case about workforce as a strategic variable, and particularly, how integrating more women into the workforce is a distinctive competitive advantage.
I like ideas like I like people, and I want to translate them. I’m interested in taking on a multi-disciplinary approach to think about innovation from questioning and developing a thesis on this issue. Specifically, I’m curious about what we can learn regarding innovation by analyzing an adversarial geopolitical situation involving technology, economics, and policy choices.
I want to know how each country approaches utilizing women in a male dominated field, and how that has second order effects for the country’s economics, innovation capabilities, and technological achievements.
Section 1: I am the law - the regulatory & legal policies surrounding China and the United States’ approach to military-civil work
The representation of women in defense technology is generally abysmal across the world, and China isn’t better at diversity as much as China is better at gaining access to labor through sheer ability which increases the representation of women.
Before I get into the sheer numbers, readers should understand the mechanisms and systems driving them in the first place.
Key points:
Women in China have their work contribute to both commercial and military use from day one due to the Military-Civil Fusion (MCF) which is a national doctrine encoded in law and policy permitting immediate dual-use cross-over and application of commercial work. This is completely unlike the U.S., where both constitutional constraints alongside security clearance processes with oft extended backlogs prevent such transitions between commercial and military usages.
China’s Military-Civil Fusion doctrine compels compulsive contributions between civilian tech work and defense applications, eliminating friction between industries for dual-use.
America’s constitutional constraints prevent such a set-up due to the First Amendment (compelled speech), Fifth Amendment (takings), and the Thirteenth Amendment (involuntary servitude).
Why is understanding the legal framework important? Through non-consensual coercion, China mobilizes their entire technical workforce instantly whereas the U.S. has legal and cultural barriers to compelled cooperation.
Women in China have their work contribute to both commercial and military use from day one due to the country’s Military-Civil Fusion (MCF) plan which is a national doctrine encoded in law and policy permitting immediate dual-use cross-over and application of commercial work. The MCF as a system of laws and policy decisions removes barriers between industries as well as compelling dual use development while making it a legal obligation to cooperate, license routes for defense industry work, and much more. Though there is no omnibus type of legislation for the MCF, it is through its infrastructure and policy mandates that the MCF is an obligatory and compelled nonconsensual mobilization of labor across all of private Chinese industry to share commercial development and innovation with the Chinese government.
This is completely unlike the U.S., where both constitutional constraints alongside security clearance processes with oft extended backlogs prevent such transitions between commercial and military usages.
The MCF doctrine uses as its legal pillars China’s respective 2010 National Defense Mobilization Law and 2016 National Defense Transportation Law alongside practical policy decisions and administrative measures to integrate work across domains for dual governmental/military use. The 2010 and 2016 laws require that civilian infrastructure and resources be built to military standards and be convertible to military use in wartime. The core MCF policy goal as repeated across official documents is to “...form a pattern of deep integration that is all-elements, multi-domains, and high-efficiency.”
Beyond those legislative instruments, China publishes sectoral laws, administrative regulations, five-year plans, and “opinions” that operationalize MCF in specific fields. Some are binding laws, whereas others are official policy documents issued by the Party or State Council linked to China’s governance model. The near seamless transition this sets up is in stark contrast with the arduous and segregated procurement methodology of the U.S. as well as its security clearance situation. Where one country is speed-running, the other is stalling.
In practice, this means when Alibaba hires a software engineer (it is a company where 47% of employees are women), that engineer works on e-commerce recommendation algorithms. The same machine learning techniques predicting what products customers want also apply to intelligence analysis, predicting what information analysts need. Through state coordination mechanisms, the commercial algorithm development simultaneously informs People’s Liberation Army (PLA) capabilities. Alibaba’s engineer doesn’t need security clearances, doesn’t need to wait eight months for background investigations, doesn’t need to do anything at all except for show up to work on Monday to write code, and both Alibaba’s commercial products and China’s military intelligence benefit from her work. Any commercial development by a Chinese company has capabilities which flow directly to the Chinese military’s use via the state’s coordination and joint laboratories, as there is no ‘division’ of the work between the “defense” and “civilian” tracks. They’re all accessible for national security priorities without bureaucratic barriers unlike in the U.S.
Meanwhile, the United States faces dual-use constraints through its legal and regulatory frameworks. Firstly, we can’t replicate Military-Civil Fusion as constitutional constraints prevent it. The First Amendment’s compelled speech doctrine means the government can’t force Anthropic or Google to hand over their AI models for military applications. The Fifth Amendment’s takings clause requires just compensation if the government wants private sector technology.The Thirteenth Amendment prohibits involuntary servitude, which is why you can’t conscript software engineers into weapons development like you could from the earlier example of the Alibaba engineer. In China, her work on commercial algorithms to identify customer recommendations can be transformed into military use to identify whatever use case they deem appropriate. We can’t copy the Chinese method due to these constitutional points and we shouldn’t.
It is anathema to everything American to forsake the values these constitutional components codify.
After all, America is agentic, all-in, always on.
That’s been our story from the founding of our country and something to leverage to our advantage. Voluntary motivation drives innovation in ways different from conscription or state compulsion. Yet constitutional advantages only materialize if we actually mobilize talent including women through voluntary attraction rather than excluding them through artificial barriers. The mechanisms forming America’s legal and policy choices currently mean that even though the ethos of being proactive is there, we live in between the limbo of being agentic (through enabling innovation) and being administrative (through bureaucratic backwaters).
In additional to the legal frameworks of constitutional constraints, regulatory measures abound as America maintains a security clearance system that’s basically economic sabotage disguised as national security. The average amount of time a Top Secret clearance takes to process is 249 calendar days. That’s eight months of waiting while China’s engineers start contributing on day one. With massive labor shortages and over 70,000 defense positions currently sitting unfilled resulting in $14 billion in foregone annual economic activities, 300,000 current cases are backlogged in a security clearance system that cost $1.7 billion to modernize and yet still doesn’t work.
Given the legal mechanisms undergirding what is possible in China with the fusion of military and civilian technologies and how that is not the case in the U.S., the labor market pipeline and its measurable outcomes serve as entry points defense analysts track quarterly.
Section 2: Gender arithmetic/the numbers situation
Here I’ll establish the competitive reality of the current landscape by comparing the raw numbers of women in tech and defense between China and the U.S., showing how the representation gap compounds at every transition point.
Key points:
China: 2.85 million women in STEM R&D (40-46% of workforce), all accessible for defense through MCF from day one, with 1 million new STEM graduates joining the Chinese labor market each year.
U.S.: Women earn 53% of STEM degrees but drop to 23-29% in defense contractors
40% of women with engineering degrees in the U.S. leave the field entirely within years (vs 10% of men)
At current rate of change (0.15% annually), Lockheed reaches gender parity in 2189
If U.S. defense matched 47% labor force gender distribution = 275,000 additional women working (which is four times the existing 70,000 unfilled positions)
China has 2.85 million women working in STEM research and development, which is around 40-46% of their total STEM workforce. The United States graduates approximately 400,000 STEM students annually, with women earning 53% of STEM degrees, but then hemorrhaging them from defense work.
America’s numbers tell a different story. Women represent only 15% of practicing U.S. engineers despite earning 20% of engineering degrees.
In absolute numbers, thousands of women engineers walk away from the field each year in America. From recent cohorts alone, if 33,000 women earn engineering degrees annually and nearly 40% leave within a few years, that’s over 13,000 women lost from each graduating cohort, which is 3 times more than MIT’s entire undergraduate enrollment.
Lockheed Martin hit 23% women in 2022. In 1976, they were at 16% women. That’s 7 percentage points of improvement over 46 years of stated commitment to diversity regardless of shifting political administrative machinations in the background.
At that rate of change (0.15 percentage points annually), Lockheed will reaches gender parity with the U.S. labor force in the year 2189. At this rate of change (0.15 percentage points annually), Lockheed reaches gender parity in 2189.
In 2189, I’ll be dead.
You’ll be dead.
Everyone reading this will be extremely dead.
Extraterrestrial life forms may be the first beings to witness when Lockheed Martin achieves gender parity (if Lockheed Martin is still around by then, and I mean, I’d hope the engineering quality by then reflects making contact with E.T. or his 3rd cousin or such…)
Notice how along the way of every transition how the representation of women drops? I didn’t even include the statistic where for example, 40% of women earning engineering degrees leave the field entirely within a few years of graduating. That’s not 40% of women leaving their first job, that’s 40% of women with engineering degrees exit engineering as a career forever.
If American defense matched the 47% labor force gender distribution, we’d have 275,000 additional women working in defense, which is already four times the 70,000 currently unfilled positions in the labor shortage plaguing the industry. We are clearly not lacking the talent, but rather, we’re excluding the talent through artificial shortages and administrative burdens.
China is adding 1 million women with STEM backgrounds into their labor market every year. We don’t even have 10% of that number being added annually (as of current writing).
China’s number of women working in STEM means approaching gender parity in fields that determine whether you build hypersonic missiles before your adversary does, whether your AI systems close capability gaps in 13 months instead of 13 years, or whether your drones dominate the skies at 10x the scale of your competitor’s fleet.
How do you engineer the best Jenga tower? You find the best talent and you nurture it. When women represent 40-46% of your technical workforce, you’re not scrambling to fill positions or delaying weapons programs because you can’t find qualified personnel. You just hire the qualified people who exist, and half of them are women, and your Jenga tower keeps growing while your competitor explains to shareholders why the hypersonic missile program is delayed another 18 months due to “workforce constraints” and a lack of access to the other half of the Jenga box pieces.
Through MCF, women contribute to defense innovation from day one. China isn’t improving representation but is approaching gender parity in the fields that determine whether you build hypersonic missiles before your adversary does, whether your AI systems close capability gaps in 13 months instead of 13 years, whether your drones dominate the skies at 10x the scale of your competitor’s fleet.
Section 3: What China’s utilization of women actually produces for measured outcomes
Here I’ll connect the dots between how more women in the workforce and their hard data capabilities reflects what gender parity produces at scale.
Drones: China fields 10x fleet size of US/Taiwan combined, DJI has 70-90% global market, 44-62% women in electronics manufacturing
AI development: US lead collapsed from 103 points to 23 points in 13 months (what analysts thought would take 5-7 years)
2 million data annotators training China’s AI (majority women in rural poverty programs)
Weapons acquisition: China acquires high-end systems 5-6x faster than US
Specifically, I’ll analyze measured outcome dynamics via drones, AI development velocity, and weapons acquisition speed as it ties to the participation of women in the labor force. This will then identify which questions matter for strategic competition in the future when one country is using all the Jenga pieces and the other has voluntarily left the other half in the box.
Drones: 10x fleet size advantage
A colorful fact that is arguably as memorable as Andruil founder Palmer Luckey’s Hawaiian shirts is that China’s drone fleet is estimated at 10 times the combined size of US and Taiwan forces. In 2024 alone, they added 720,000 drones. DJI commands 70-90% of the global consumer drone market. Where do women come in? Well, here’s the gender component that matters - China’s electronics manufacturing workforce is 44-62% women depending on the sector. These women assembling consumer devices and military components aren’t in separate facilities. It’s the same factories with the same workers and the same production lines.
This production integration means China can surge military drone production by redirecting existing capacity without building new facilities, hiring new workforces, or waiting for security clearances. Meanwhile, per PricewaterhouseCoopers, US defense contractors and their tens of thousands of unfilled positions translates into how they cannot scale drone production because they cannot find enough cleared personnel to staff production lines.
AI development velocity
In January 2024, US AI models held a 103-point performance lead over Chinese equivalents on standardized benchmarks. In February 2025, just thirteen months later, that lead had collapsed to 23 points. China closed 80% of a capability gap that analysts expected would take five to seven years to bridge. They did it in thirteen months. When it comes to production timing, China closed that gap faster than Lana del Rey’s most prolific year in music.
Outside of Lana’s later life choices, here’s the part that should really give you pause and drive the point home for you in case everything else before didn’t: remember how I said China employs approximately 2 million data annotators training their AI systems—the majority are women, particularly middle-aged rural women in poverty-alleviation programs. In Qingjian County, one of China’s poorest regions, the Aidou project employs 70% mothers from the 1990s generation, processing 20,000+ images daily for 4,000-5,000 yuan monthly—among the best-paying local jobs. These women are literally training the AI models that the PLA uses for battlefield scenarios, facial recognition, and autonomous weapons. China turned rural female poverty into a strategic AI advantage. All of Chinese society is benefitting massively from the winds of technological change and their output is reflecting this warspeed reality.
Weapons acquisition speed
A 2024 analysis concluded that China acquires high-end weapons systems 5-6 times faster than the United States. Five to six times. Not 10% faster or 25% faster. An order of magnitude faster. Pentagon assessments describe the PLA as operating on “wartime footing” versus America’s “peacetime footing.”
When you can immediately staff 10 parallel development programs because you’re drawing on 2.85 million women in STEM R&D who can start work instantly without clearance delays, you move faster than a competitor who can only staff 2-3 programs due to an artificially constrained workforce. It’s not as if the work is better or worse, if people are working harder, etc. It’s just the numbers. Literally. Much like dating, much like job hunting, much like engagement farming, much like anything, it’s a numbers game. More teams testing more approaches equals faster identification of what works. China understood the assignment. We have too, but to what pace?
Section 4: American amnesia and next steps:
Today is November 10th, 2025, and as of writing, to recap with clarity, here is where we are:
China has 2.85 million women in STEM R&D, which is 40-46% of their technical workforce.
All of them are accessible for defense work from day one because the MCF doctrine means there’s no separation between “commercial” and “military” work. It’s just work and all of it counts.
In the United States, we graduate women with 53% of STEM degrees and then proceed to lose 40% of them from engineering as a subfield forever.
Our major defense contractors sit at 23-29% women while 70,000 positions stay empty because we can’t figure out how to staff them with security clearance bottlenecks left and right.
China fields 10x our drone fleet. They closed an 80% AI capability gap in 13 months when our analysts said it would take 5-7 years. China acquires weapons systems 5-6x faster than we do.
I’ve mentioned Jenga multiple times in this post, and it’s because the introductory backstory about it is inspired by a national security conference I attended alongside memories of a game I play with a Chinese friend. Go read the post. Come back to this one. Jenga is the symbol for defense technology, in case it wasn’t clear.
China isn’t inherently better at playing Jenga, it’s just because they’re using all the pieces and we’re leaving half of ours in the box.
We’re not falling behind and losing because our framework is wrong. We’re losing because we’re voluntarily restricting our own talent pool including women for no strategic reason whatsoever.
Our competitive reality in defense technology currently reflects a cognitive disassociation between the endless march of American domination and dynamism without extending any dignity to our inherent advantages as a nation of agentic individuals - China forces its women to participate while we exclude ours by artificial designs and barriers. There is no dignity in being dominated and there is no dignity in domination. There is only success - and the reality is that China is far more successful at accelerating innovation in defense technology than we are right now in large part due to their utilization of women and all the practical as well as innovation work this contributes forward. The wild part - beyond the facts and figures of our current state of things - is that we’ve actually done this before in America where we mobilized women in defense technology, and the results were decisive enough to change the outcome of World War II, and then we just... forgot. China, on the other hand, remembered. We forgot. Part 2 examines what we learned, what we forgot, and what that amnesia has cost us.