DeepSeeking the Truth

Last December, the world knew little of DeepSeek. But earlier this year, the Chinese startup drew international attention after it released DeepSeek V3, an artificial intelligence model that matched Open AI’s model and outperformed then leading open-source models, including from Meta. The real breakthrough, however, was that the company trained its model with under $6 million—a fraction of the amount that other firms have spent—and without access to the most advanced chips. DeepSeek’s feat recalibrated investor’s assumptions about AI hardware dependencies, driving the chip manufacturer Nvidia’s stock down by 17 percent and vaporizing nearly $600 billion in market capitalization. As CNN put it, this bootstrapping little startup posed a threat to “the aura of invincibility surrounding America’s technology industry.”

Many panicked commentators in the West hastily framed it as a product of the Chinese state, assuming it benefited from government backing, industrial espionage, or a combination thereof. OpenAI’s Sam Altman called DeepSeek “state-controlled,” and tried to rally regulators to ban “PRC-produced” models. Others dismissed its achievements outright, arguing that Chinese AI companies merely copy and refine innovations that originated in the United States. These narratives flatten China’s innovation landscape into a one-dimensional story, but the reality is more complex: DeepSeek was neither the result of a top-down state initiative nor built on stolen technology. Instead, it emerged from the fringes of the country’s tech sector, funded by private capital and driven by a small, independent team. While its success may give the impression that China has secured a strong foothold in AI’s future, it would be premature to claim that DeepSeek validates China’s broader innovation system. Its success came about not because of that system but, in many ways, in spite of it.

In an era of rapidly intensifying tension with China, our understanding of Chinese innovation has serious geopolitical stakes. Misreading cases like DeepSeek risks overestimating China’s tech sector and flattening our understanding of the world’s second-largest economy. And at a time when American firms are investing billions to secure domestic AI advantage, a clearer grasp of what enables—or constrains—breakthrough innovation in China is essential. DeepSeek offers a rare window into that complexity. Its story invites us to rethink not just how China innovates, but what kinds of labor systems, organizational cultures, and institutional arrangements actually foster technological breakthroughs in the era of AI.

DeepSeek’s success is not just a product of technical ingenuity but also deeply rooted in its unique approach to labor relations, which runs contrary to the grueling work schedules, rigid hierarchies, and relentless internal competition that have become the norm across China’s tech industry. Perhaps the most notable aspect of the industry is its long-practiced “996” work regime: 9 a.m. to 9 p.m., six days a week. During those hours, employees are kept on a tight leash, subject to stringent reporting requirements, assessed by key performance indicators, and regularly ranked against one another. Since the mid-2010s, these long hours and draconian management practices have been a staple of the industry, considered the bare minimum required to catch up to Silicon Valley.

Indeed, speed was paramount during China’s digital growth years, when companies were focused on aggressive user growth and market expansion. For workers, this meant quickly and continuously rolling out new features and products to outpace competitors and capture market share. This relentless pursuit of expansion demanded a workforce that functioned like a well-oiled machine; each employee a disposable foot soldier in the battle for market share.

This relentless workplace culture echoed the early days of Silicon Valley, where workers were known to far exceed the standard forty-hour workweek, if not sleep in the office. But by the 2010s, labor relations in the U.S tech sector began to shift. Recognizing that the structure of work could be more valuable than any specific line of code that the company owned, companies like Google and Facebook gave employees tremendous autonomy and spent lavishly on perks and high salaries. Despite efforts to keep employees tethered to the office, average working hours declined: by 2024, more than 30 percent of tech workers reported working fewer than forty hours per week. That said, while tech workers in the United States have had good working conditions, it’s far from a utopia. Workers have butted heads with their employers over the uses of the technology they build—for instance, workers at Amazon and Google have pressured their employers to stop working with U.S. border and immigration agencies, to no avail. And across the industry, unionization efforts have gathered steam, but few have mounted to much. And since 2023, amid a wave of layoffs, companies have tightened control over workers.

Meanwhile, Chinese giants like Alibaba and Tencent made the 996 work schedule the standard for labor relations across China’s tech sector—at least until the early 2020s. Management is known to use workplace communication platforms like DingTalk—not just for internal messaging but for its built-in surveillance functions. For example, one of the app’s most popular features is its digital punch card which can track an employee’s location. The app also measures employee productivity. Those who fail to meet performance benchmarks risk demotion, loss of bonuses, or even termination, leading to a culture of fear and relentless pressure to outperform each other.

Since the late 2010s, however, China’s internet-user growth has plateaued, and key digital services— such as food delivery, e-commerce, social media, and gaming—have reached saturation. While these developments have put the efficacy of this model under strain, tech firms have doubled down on squeezing their workforce and cutting costs. Rather than improving their technology with research and development, they’ve focused on applying it to new business models that leverage network effects in a short-sighted effort to maintain or expand their market share. This approach comes at a cost: stifling China’s ability to engage in long-term innovation-based competition.

DeepSeek’s approach represents a marked departure from industry norms. Since its founding in 2023 by entrepreneur and hedge-funder Liang Wenfeng, the company has built a workplace culture centered on flat management, academic-style collaboration, and autonomous young talent. The workforce is kept deliberately small, at about 150 employees, and management roles are de-emphasized. Based on numerous reports in the Chinese media, DeepSeek’s teams are formed around specific goals, with no fixed hierarchies or prescribed roles. Team members focus on tasks they excel at, collaborating freely and consulting experts across groups when challenges arise. This approach enables ideas with potential to receive the appropriate resources. One twenty-six-year-old employee said that “Liang gave us decision-making power and treated us as experts. He constantly asked questions and learned with us.” Other employees have reported that Liang does not enforce the kind of top-down management seen in other tech companies. There are no weekly reports, no forced competition, no KPIs, and, famously, no seventy-two hour workweeks.

DeepSeek has focused its recruitment efforts on young but promising individuals over seasoned AI researchers or executives. Many of its researchers, including those who contributed to the groundbreaking V3 model, joined the company fresh out of top universities, often with little to no prior work experience. As one headhunter who worked with DeepSeek told a Chinese media outlet, “They look for three to five years of work experience at the most. Any more than eight and you’re just a ‘pass’ for them.” Liang explains the bias toward youth: “Real innovation often comes from people who don’t have baggage.” While many Chinese tech firms prize younger candidates for their willingness to work longer hours, state-backed firms like Zhipu often seek out high-profile seasoned industry recruits to bolster credibility and drive tech transfer from incumbents.

DeepSeek’s funding and corporate-governance structure is equally unconventional. Unlike China’s electric vehicle, battery, and solar industries, DeepSeek didn’t rely on state-backed initiatives or investments from tech incumbents. Instead they sought investment from a hedge fund operated by Liang, which also acquired ten thousand NVIDIA chips for DeepSeek to get off the ground. Where most Chinese tech firms are under constant pressure to chase short-term profits, appease investors, or align with state-defined KPIs, DeepSeek’s unique corporate governance provides it with access to patient capital, allowing it to pursue longer-term, riskier research agendas without the immediate imperative to commercialize.

DeepSeek’s success highlights that the financing, work, and incentive structure underpinning technological development are critical for innovation. While many of China’s tech giants have focused on squeezing diminishing returns from overworked employees, DeepSeek has demonstrated that by breaking away from the hierarchical, control-driven norms, it can achieve results that outstrip state-funded competitors like Zhipu or better-funded corporate competitors like, Tencent, ByteDance, and Alibaba.

This brings us to the larger question of how DeepSeek’s success fits into ongoing debates about China. In the press, the debate often flip-flops between two starkly opposing and equally flawed views: China is either doomed or it is the next technology superpower and is thus a pressing threat to the United States. As I see it, this divide is about a fundamental disagreement on the source of China’s growth: whether it relies on technology transfer from advanced economies or thrives on its indigenous ability to innovate. Though these are often presented as mutually exclusive positions, this is somewhat of a false dichotomy. They are instead part of the same sequential progression—technology transfer laying the foundation for the emergence of homegrown innovation. Still, it’s useful to distinguish between the two approaches.

Those who believe China’s success depends on access to foreign technology would argue that, in today’s fragmented, nationalist economic climate—especially under a Trump administration willing to disrupt global value chains—China faces an existential risk of being cut off from critical modern technologies. From this perspective, isolation from the West would deal a devastating blow to the country’s ability to innovate. On the other hand, those who believe Chinese growth stems from the country’s ability to cultivate indigenous capabilities would see American technology bans, sanctions, tariffs, and other barriers as accelerants, rather than obstacles, to Chinese growth. As the tech analyst Dan Wang wrote in 2019: “The US responded to the rise of the USSR and Japan by focusing on [domestic] innovation . . . the US is responding to the technological rise of China mostly by kneecapping its leading firms. So instead of realizing its own Sputnik moment, the US is triggering one in China.” In this view, such restrictions compel Chinese firms to innovate, upgrade, and develop homegrown technological solutions, ultimately strengthening self-reliance and long-term competitiveness.

So far, this debate has primarily unfolded in the context of advanced manufacturing sectors, from solar to batteries, and, more recently, electric vehicles. For years, the prevailing theory was that Chinese firms needed to first acquire fundamental technologies from the West, leveraging this know-how to scale up production and outcompete global rivals. So the initial restrictions placed on Chinese firms by the first Trump administration, unsurprisingly, were seen as a major blow. For instance, in 2018, after Trump banned one of China’s largest electronic makers ZTE from using U.S.-made components, the company’s operations effectively grinded to a halt, with the New York Times calling the sanction a “death sentence.” China’s continued dominance in solar, batteries, and electric vehicle production, however, has shifted the narrative to the indigenous innovation perspective, with local R&D and homegrown technological advancements now seen as the primary drivers of Chinese competitiveness.

The success of China’s digital/high-tech sector, however, is usually seen as dependent on technology transfer. Part of the reason is that AI is highly technical and requires a vastly different type of input: human capital, a front in which China has historically been weaker and reliant on foreign networks to make up for the shortfall. Scholars like MIT professor Huang Yasheng attribute the rise of China’s tech sector to the many collaborations it has had with other countries. Even Chinese AI experts think talent is the primary bottleneck in catching up. Indeed, China’s post-2000s information and technology sector built its success on the back of overseas technical know-how. Many of China’s early tech founders either received education or spent considerable time in the United States, and they would go on to privilege employees with overseas experience, particularly those who have worked in U.S.-based tech firms. In the generative AI age, this trend has only accelerated: Alibaba, ByteDance, and Tencent have each set up R&D offices in Silicon Valley to increase their access to U.S. talent.

This is where DeepSeek diverges from the traditional technology transfer model. Instead of relying on foreign-trained experts or international R&D networks, DeepSeek exclusively uses local talent. Liang himself also never studied or worked outside of mainland China. The DeepSeek story shows that China has the indigenous capacity to push the frontier in LLMs, but needs the right organizational structure to flourish. Much like China’s advancements in solar manufacturing, batteries, and electric vehicles, DeepSeek symbolizes a critical turning point in AI and tech writ large. The real test lies in whether the state-supported ecosystem can evolve to nurture more companies like DeepSeek, or whether such firms will remain rare exceptions. As of now, the latter seems more likely.

Among the tech elite of the West, DeepSeek’s rise challenges a deep bias: the jingoistic assumption that any major Chinese tech breakthrough must be the product of state support or industrial espionage. This characterization of DeepSeek also reflects a deeper racialized disbelief in Chinese ingenuity. When Sam Altman casually referred to DeepSeek as “PRC-produced” AI, for instance, he wasn’t just wrong. These kinds of claims serve a useful purpose for the American tech elite. By framing Chinese firms like DeepSeek as state-supported threats, tech leaders create the case for greater state support at home. It’s a strategic narrative: Big Tech using the specter of China to secure investment and prevent regulation. In this light, accusations about DeepSeek’s origins aren’t just misinformed—they’re revealing. They tell us less about China and more Silicon Valley’s anxiety and how it plans to reshape the state in its own image.