In the early 1980s, a cluster of fledging computer companies opened up shop in a chaotic corner of northwest Beijing, near the campuses of Peking and Tsinghua Universities. Electronics Street, as the area became known, was a tangle of sturdy bicycles and hand-drawn signs, loud with heated bouts of haggling. Dusty banners hung over pedestrians’ heads, while boxes of copy paper stacked 10 or 12 high blocked their path. Men in cheap suits hawked power strips and printer ink from street-side kiosks. Pirated software was so abundant that some preferred the moniker Crook Street.

The existence of a burgeoning PC market was remarkable, given that many Chinese still did not own a refrigerator. But more remarkable was that the businesses of Electronics Street were private enterprises. Their foray into capitalism was an experiment launched with China’s economic reforms, which early on were linked to investments in science and technology. Early signs suggested that this strategy might just work. Among the companies to emerge from the detritus of Electronics Street was Lenovo.

In the years since, China has undergone a scientific and technological renaissance. Between 1991 and 2016, government funding for research and development grew by a factor of 30. The country overtook Japan in spending on R&D back in 2009. The Organization for Economic Cooperation and Development predicts that it will outspend the United States by 2019. Today Electronics Street is known as Zhongguancun, and it is home to the tech giants Baidu, Didi Chuxing, and Meituan-Dianping, along with research centers for Microsoft, Google, and IBM.

China boasts nine of the world’s 20 largest tech companies, including three in the top 10. The country can claim the world’s largest single-dish radio telescope and several of the fastest supercomputers, and it plans to build the largest supercollider. In 2016, it launched the world’s first quantum communications satellite. The ambitions charted in recent government plans are far-ranging: to excel in areas like 5G mobile technology, seed breeding, and robotics by 2020 and to become a world leader in artificial intelligence by 2030.

All of this has prompted considerable anxiety in the United States. Citing concerns about mercantilist market controls and industrial espionage, the Trump administration has plunged into a trade war with China. In October, Vice President Mike Pence accused the Chinese government of perpetrating “the wholesale theft of American technology.”

To some extent, American talk about a technological cold war conceals a broad perception gap. While US lawmakers see in China’s scientific aspirations an acute threat, Chinese commentators see a lingering insecurity. To them, the ambitions staked out on Electronics Street have not been fully realized. Government white papers and state press reports may project bravado, but in more intimate settings Chinese leaders lament that the country has spent much and gotten little to show for it. Yes, China funds big science projects, but that is not the same as achieving serious scientific breakthroughs or launching a product that reshapes the global market, like the iPhone. Even as China’s elite universities climb the world rankings, the country can claim only one Nobel laureate in the sciences who did not do her prize-winning work abroad.

Yet there are signs that this may be changing. While a wave of homegrown Nobels may still be a way off, the country has seen an explosion of business innovation. The country’s powerful tech companies, along with a few ambitious startups, are now shaping business models in Silicon Valley—and driving debate over internet controls and surveillance in the process. They have succeeded in large part because of a scrappy entrepreneurialism of the sort first seen on Electronics Street. As they get bigger and set their sights overseas, what holds them back may no longer be a lack of talent or resources. Rather, it may be their ties to the Chinese government—the very institution that set off China’s tech boom when it began the economic reform program 40 years ago.

For years scholars asked: How can China innovate when science and technology are administered top-down? How can researchers achieve breakthroughs when there is no free speech, limited freedom of inquiry, not even access to Google Scholar?

Zhongguancun seemed like a telling example, at least at first. In 1989, some of the Electronics Street entrepreneurs joined the Tiananmen Square protests, which first took shape at nearby Peking University. When the crackdown came, Communist Party cadres were dispatched to

Zhongguancun to coerce techies into denouncing the democracy movement. By the end of the 1990s, the area was formalized as a science park under more direct municipal government supervision.

Zhongguancun was deemed “China’s Silicon Valley,” a comparison that was absurd from the start. The top-down approach implemented in the years that followed is a far cry from the more decentralized innovation of the San Francisco Bay Area. It grew more absurd as the Zhongguancun model was replicated throughout China 167 times. By the early 2000s, many of these science parks struggled to attract high-caliber tenants. Some became mere distribution and processing centers for foreign technology firms.

The notion that simply populating China with science parks would lead to progress reflected a government emphasis on quantity and metrics at the expense of quality. Consider the Chinese technology plans. One key plan, adopted in 2006, mapped out growth targets to 2020. By that year, China aimed to spend 2.5% of its GDP on R&D and excel in areas like biotechnology, nanotechnology, and drug development. Other goals included making China one of the top five countries in the world in both invention patents and total number of citations.

Some of these targets were reactive: the inclusion of nanotech followed the establishment in 2000 of the US National Nanotechnology Initiative, which provides more than a billion dollars a year in funding for nanoscience research. But they got a Chinese spin. After the 2006 plan was unveiled, local governments rushed to show their support by unveiling incentive schemes. Professors’ salaries depended on how many papers they published in indexed journals. For companies, lucrative innovation subsidies hinged on getting large numbers of patents. Both the national and provincial governments devoted significant sums of cash to attracting back the tens of thousands of Chinese-born researchers living abroad, reasoning that they might jump-start innovation. Grant recipients were offered lucrative resettlement sums, along with salaries far above local norms.

The effect was dramatic, at least on paper. Output skyrocketed. Today China ranks first in the world, ahead of the United States, in the number of science and engineering papers published in international journals, according to the US National Science Foundation. The World Intellectual Property Organization places it second in international patent filings. In AI, one of the government’s key strategic areas, China leads the world in both published papers and issued patents. It also leads in nanotech patents, according to an analysis by Tsinghua University’s China Institute for Science and Technology Policy.

But look closer at these numbers and problems appear. Many domestic Chinese patents are so-called junk patents that are not renewed after their fifth year. The emphasis on publishing has led to a flourishing black market in academic publications. Shady agencies hawk ghostwriting services on the messaging app QQ, while enterprising researchers sell authorship spots on papers that have been accepted to indexed journals. Mainland authors are second only to those from US institutions in the number of English-language papers churned out every year, but many make little impact on global scholarship. Globally, the average English-language paper is cited 11.8 times; for scholars from Chinese institutions, the figure drops to 9.4 times.

A side effect of the enthusiasm for metrics is that less measurable areas, like teaching and mentoring, have suffered. Researchers are rewarded only for first and last authorship slots on papers, so many see little point in collaborating on large projects. Around the world, incentives often discourage work on big research questions requiring years to answer, but in China this is especially true. Bai Chunli, president of the Chinese Academy of Sciences, recently complained that China needs more scientists willing to “take a decade to sharpen a sword.”

As spending on R&D has increased, a portion of the money has been spent on luxury cars, bribes, and mistresses. In one city in Guangdong province, science administrators reportedly pocketed 30% of the grant monies they handled. Money allocated for luring back scientists from overseas has been wasted as well. In Washington, Congress is worried about the Thousand Talents Plan, a recruitment program targeting foreign and overseas Chinese researchers. But Chinese leaders have a very different concern. Most of the Chinese scientists and entrepreneurs recruited under the program return to China only for part-time gigs and make little lasting contribution to research in the country. Some cynically see the grants as a sort of vacation plan—a way to fund visits to family and friends while keeping their tenured posts in the West.

One reason overseas Chinese scientists are reluctant to move back home is cronyism. With the Chinese Communist Party still firmly in control of most universities and research institutions, researchers are expected to spend time wooing administrators. Graduate students and postdoctoral scholars know that if they make a breakthrough discovery, their supervisor might swipe it.

And yet, despite all this, some areas of science and technology in China have flourished. Why? Critics like Pence are quick to point to China’s absorption of foreign technological know-how through joint ventures, open-source collection, and blatant industrial espionage. Others cite its growing consumer population, which offers startups the prospect of instant scale. Certainly, both have played a role.

But there is another, less obvious reason for China’s rise: the interaction of state directives with grassroots experimentation—or what a new book, Innovation in China, by Richard P. Appelbaum and several coauthors, calls the “often contradictory blend of heavy-handed state-driven development and untrammeled free enterprise.” This is most evident with Chinese internet companies.

The leading tech companies Baidu, Alibaba, and Tencent are often referred to by the acronym BAT. That term hides a regional and entrepreneurial breadth. Tencent is headquartered in Shenzhen. Alibaba and its affiliate Ant Financial are in Hangzhou. Only Baidu is based in Beijing’s Zhongguancun. The major thing the three companies have in common is that they have benefited from varying levels of blocking, throttling, censorship, and other measures that have hobbled foreign competitors in a market hungry for internet products.

A number of early Chinese sites were overt clones of censored American ones. Technical problems reinforced their advantage. Web traffic at the time was filtered through the Great Firewall in only three locations: Beijing, Shanghai, and Guangzhou. When foreign sites weren’t blocked, or when users had access to workarounds, they loaded at sloth-like speeds. In the time it took to load a YouTube video summarizing War and Peace over a VPN, it was almost possible to read the book.

But the copycat sites were not anointed winners, at least initially. They battled it out with a pool of ruthless competitors “like gladiators in the coliseum,” in the words of Kai-Fu Lee, a venture capitalist and former head of Google China. Fresh-faced entrepreneurs had the stamina to tweak their products until the interface and features were what Chinese people wanted—which often differed significantly from Western norms. They intuited that certain holidays and other cultural phenomena could be spun into blockbuster promotions. Alibaba, for example, used a national fascination with numbers to transform November 11 into the online shopping blitz Singles’ Day, which now dwarfs the US’s Black Friday in sales. They were also willing to resort to underhanded tactics to win market share. In his book AI Superpowers, Lee tells how the CEO of Renren, a Facebook wannabe, bought a URL that could have been mistaken for that of competitor Kaixin001, copied the user interface, and signed up his rival’s customers. Kaixin001 was run out of business.

Most important, China’s internet startups were willing to try out ideas that hadn’t been proved elsewhere. One such idea was the integrated QR scanner—the technical tool that enables mobile payments within the apps WeChat and Alipay. QR codes link online and offline worlds in unexpected ways, turning basic objects like menus or bus schedules into a virtual experience and making it possible to move through a city with the constant aid of technology. Alipay introduced QR codes in 2011, a full year before US technologists unsuccessfully hyped them at the Texas tech festival SxSW. WeChat soon followed suit.

The scanners are a big reason that China’s mobile payments market is now valued at $ 15.4 trillion, making it more than 40 times as large as the US one. That market underlies one of the fastest-growing sectors of the Chinese tech world: online-to-offline services. Startups enable consumers to order house calls from tutors, hair stylists, and dog groomers at a moment’s notice. Companies like Ofo and Mobike have flooded cities in China with millions of dockless bikes, transforming public transportation. These new services are not without kinks—Chinese cities have struggled to deal with massive bike-share graveyards—but they have made the lives of many middle-class Chinese infinitely more convenient. Megacities no longer feel so daunting when you know that you can farm out unpleasant daily tasks, or hop on a bike if you can’t find a taxi.

Chinese tech companies have done so well at this sort of experimentation that the direction of copying has reversed, with US tech companies now borrowing ideas from the mainland. The CEO of Kik, Ted Livingston, has said he aims to make the app the “WeChat of the West.” (He got $50 million in financing from Tencent.) Ofo and Mobike have expanded throughout the world, inspiring copycats everywhere—though they have foundered in less densely populated cities where mobile payments aren’t so popular.

It is this frenetic energy and intense competition, not Chinese government attempts to pick winners and set targets, that is driving innovation in China. Across sectors, the most exciting companies began as renegade startups. Genetic-research powerhouse BGI spun off from the Chinese Academy of Sciences in Beijing, but later moved to freewheeling Shenzhen. Drone maker DJI was founded by a university student working out of his dorm room in Hong Kong. The speech-recognition firm iFlytek was started by a group of PhD students in Anhui province. Small and medium-sized enterprises produce 80% of China’s most innovative products, according to a World Economic Forum white paper.

The first wave of Chinese innovation is in business models, not the technological breakthroughs targeted in government white papers. But one may drive the other. As they mature, China’s tech giants are opening research facilities overseas and focusing on areas like AI and self-driving cars. Baidu, which aims to have half the users of its maps app come from outside China by 2020, has two research facilities in the Silicon Valley suburb of Sunnyvale and a third in Seattle. iFlytek will open its own center in the Bay Area later this year. As these companies hire researchers from a broader variety of backgrounds and absorb ideas from abroad, they may finally have a chance at producing a world-­changing product like the iPhone.

But that’s if their ties to the Chinese government, in its growing crackdown on political opposition, don’t interfere.

Several times a year, the 25-person Politburo gathers for a “study session” on a topic of pressing concern to the Party, like market reforms or civil unrest. Typically, these are held at the Party’s Zhongnanhai complex in central Beijing. But on September 30, 2013, officials wearing matching windbreakers, dark trousers, and sensible leather shoes boarded a bus with tinted windows and rode to Zhongguancun for the first session ever held off site. There, Xi Jinping gave a speech about technological ascendance. “We must seize the opportunities presented by this new phase of technological revolution and industrial change,” he avowed. “We cannot wait, we cannot watch from the sidelines, we cannot slack off.”

The next year, the government unveiled a push for “mass innovation.” Premier Li Keqiang announced that startups would be nurtured and incubators established throughout China. For years, grassroots experimentation survived without much state encouragement, and in some cases despite strong-armed intervention. Now, finally, it is being supported.

Zhongguancun has gotten a $1.5 billion face-lift. The vestiges of Electronics Street have been cleared out to make way for shared workspaces, incubators, and startup offices on a pedestrian street called Inno Way.

This shift in policy is an encouraging development. But if China’s tech startups are now nurtured, they are also increasingly co-opted. Unlike the internet or free scholarly inquiry, mobile payments and facial recognition do not threaten authoritarian rule; they reinforce it. WeChat payment data can reveal where a person went on a particular day, down to the minute. An advanced facial-recognition system can tell where that person is right now. Since he took office in 2012, Xi Jinping has moved to swiftly consolidate power, stifle dissent, and persecute China’s Uighur Muslim minority. The tech companies have been enlisted in that quest.

According to Human Rights Watch, iFlytek is helping the Chinese government develop a national biometric database for voice recognition, with the goal of identifying speakers in phone conversations. (iFlytek has a research collaboration agreement with MIT’s Computer Science and Artificial Intelligence Laboratory.) The company Leon Technology manages surveillance infrastructure in Xinjiang, where as many as 10% of adult Uighurs are forcibly held in internment camps. SenseTime, one of the world’s leading makers of face-recognition software, has set up a joint venture with Leon Technology. Tencent and Ant Financial are piloting electronic national IDs that are embedded inside WeChat and Alipay. Ant Financial enforces government blacklists by penalizing the blacklisted in its Zhima Credit program. In 2017 BGI launched a genetic testing center in Xinjiang, where authorities are collecting DNA from Uighurs.

Such alliances could foil the companies’ efforts to create breakthrough products that appeal to consumers around the world. Worries about surveillance have already blocked Huawei’s phones from the US, though it still ships more of them worldwide than Apple. If Chinese tech companies face more concerns from overseas, they might try to distance themselves from the more nefarious government projects.

The effort to please government interests could also stifle innovation in a more fundamental way. As product life cycles accelerate, companies need to be agile to keep up. It is hard for them to do so when they have to gauge policy direction in a state that is still fundamentally opaque.

Still, it would be a mistake to underestimate the Chinese tech firms’ resourcefulness. Bets against the emergence of innovation in China have so far been losing ones. That may remain true even as the state tightens its grip.

Mara Hvistendahl is a US-based science journalist who spent eight years in Shanghai.