The most consequential bilateral technology relationship in the world has gone sour, with China and the United States lobbing tariffs, slapping procurement bans, and imposing investment restrictions. There has been plenty of rhetorical bombast on both sides, but behind the public posturing are two highly influential policy documents.
Most notable of these is the United States Trade Representative’s (USTR) Section 301 report. As I argued previously on MacroPolo, that 180-page document laid out a compelling account of China’s attempts to acquire US intellectual property, and formed the basis for a looming World Trade Organization complaint as well as potential investment restrictions and tariffs.
But another US government-commissioned report has also played a major, though less prominent, role in current tech tensions: the DIUx report on Chinese venture capital (VC) investment in Silicon Valley, titled “China’s Technology Transfer Strategy: How Chinese Investments in Emerging Technology Enable a Strategic Competitor to Access the Crown Jewels of US Innovation” (pdf here). DIUx stands for the Defense Innovation Unit Experimental, an organization under the Department of Defense (DoD) that contracts with commercial technology providers to build tools for the DoD. The DIUx report was initially circulated internally in early 2017, and was finally made public in March of 2018.
Long before it reached public eyes, the DIUx report was already having an impact on China-related legislation: Senator John Cornyn (R-TX) cited it as an inspiration for his 2017 bill that would greatly tighten scrutiny on Chinese technology investments. That bill proposed increasing the scope of deals reviewed by the Committee on Foreign Investment in the United States (CFIUS), including for the first time early stage VC investments targeted by the DIUx report.
So what is in the DIUx report itself? And does it make a strong case that Chinese VC investments are part of a larger state-sponsored technology transfer strategy, one that threatens to erode the edge of the US military and national defense ecosystem?
The DIUx report does contain a prescient analysis of China’s overarching technology transfer regime, and it raises important questions about the US government’s ability to maintain its edge in emerging technologies. But where it falls far short is fully understanding the subject of its purported focus: Chinese early-stage VC investments.
By portraying Chinese VC investors as players in a technology transfer scheme, the report misunderstands the key drivers of investment and the players involved. To be sure, new controls on Chinese technology investment will likely be needed to maintain America’s technological edge. But by misrepresenting the nature of Chinese VC investments, policy solutions that derive from this inaccurate diagnosis risk harming US innovators. Even among opponents of restrictions on Chinese investment, the nature of that harm is misconstrued: it’s not a question of losing access to cash (a commodity that is abundant in Silicon Valley), but instead losing access to irreplaceable links in global innovation networks—links that are now often located in Chinese cities like Shenzhen.
At the center of the DIUx report is an undisputed fact: Chinese VC investments in Silicon Valley skyrocketed between 2013 and 2015. In that span of time, the total value of funding rounds that Chinese investors participated in leapt from $1.17 billion to $11.52 billion, according to CB Insights. In 2015, that $11.52 billion represented 16% of all VC funding in the United States that year (these figures include money from all investors involved in a funding round, meaning the actual contribution of Chinese VCs is likely just a fraction of these totals). In 2016 and 2017, the value of Chinese-involved funding rounds have been cut nearly in half to around $6.2 billion, though this total remains nearly five times the level in 2013.
The report goes on to use graphs from CB Insights that show how substantial chunks of Chinese capital in the United States have flowed into technologies such as artificial intelligence, robotics, and augmented or virtual reality. It argues correctly that many emerging technologies are fundamentally dual-use in nature, and that the center of gravity in these fields has shifted away from in-house military research and toward the private sector.
All of the above facts are without dispute, and they alone warrant an in-depth analysis of the impact of Chinese VC investments in Silicon Valley. Are the Chinese investors driving this boom doing so at the government’s bidding? Are early-stage VC investments aimed at gaining access to core technologies that are then stolen or copied? Are Chinese VCs truly making off with the “crown jewels of US innovation”?
Unfortunately, here the report largely runs out of insights into Chinese VC activity. It makes reference to a handful of China-based investors operating in Silicon Valley—Westlake Ventures, Sinovation Ventures, and Hax—and vaguely hints at “government sponsorship” for groups like Sinovation. (Disclosure: I have worked with Sinovation founder Dr. Kai-Fu Lee on a project analyzing trends in Chinese and American AI.) Beyond that listing of players, the authors make almost no attempt to illustrate how these VC investments, which do not lead to control over a company or its technology, constitute a mechanism for technology transfer.
The most helpful illustration of Chinese government ties to VC activity comes in the appendix, where the report’s authors reprint a 2017 infographic from the tech news site The Information, detailing five Chinese VC funds launched by different branches of the Chinese government. But outside of that chart—which covers some of the Valley’s least active Chinese investors—the report presents no evidence or examples of the Chinese government directing VC activity.
The remainder of the report is dedicated to outlining different aspects of Chinese technology transfer: industrial espionage, cyber theft, and numerous legal avenues, including academic study and making scientific journals available to the public. Several of these were covered more thoroughly in USTR’s Section 301 report, though the DUIx report deserves credit for drawing attention to some of these practices more than a year before the Section 301’s release.
But questions emerge when the report’s main thesis is further scrutinized: the claim that Chinese VC activity supports technology transfer. The “VC-as-tech-transfer” claim is based less on evidence than on inference. That is, since the Chinese government wants US technology, and Chinese VCs are investing in US startups, therefore, Chinese VC investments are accomplices or pawns in this grand scheme. The conclusion isn’t incompatible with the preceding premises, but inference is hardly proof, and the conclusion seems implied rather than based on evidence. To determine whether that conclusion is warranted, let’s take a closer look at the VCs in question.
DIUx name-checks a handful of VCs, but singles out two for closer examination in the index: Sinovation Ventures and Hax. The report describes Sinovation as a “great example” of an active Chinese VC in Silicon Valley. As evidence of Sinovation’s “government sponsorship,” the report cites awards given to the firm, including a “National-Level Technology Company Incubator” designation granted by China’s Ministry of Science and Technology, as well as similar honorary titles granted by the Beijing municipal government.
Designations like these are a dime a dozen in China, applied to everything from water parks to temples. They are a far cry from evidence of government control over an entity. Since the report’s release, Sinovation’s “AI Institute” in Beijing partnered with China’s Ministry of Education to create a crash course in AI for university professors. But again, that partnership on an AI course (one that includes lectures from leading American and Canadian researchers) does not demonstrate collusion in technology transfer.
DIUx’s depiction of Hax is even more baffling. Hax is described as a “Chinese venture firm,” one that has participated in nearly half of all deals involving Chinese investors from 2014-2016. In fact, Hax is a hardware accelerator founded and run by a team of European entrepreneurs, and financially backed by Sean O’Sullivan’s SOSV, a global VC firm that invests in accelerators and their startups. The report mistakenly identifies SOSV as a venture firm “with headquarters in Shenzhen and an office in San Francisco,” which it seems to have copied directly from Hax’s description.
In practice, Hax brings startups (primarily from the United States and Europe) to Shenzhen for a 4-8 months crash course in prototyping and manufacturing their hardware products. The accelerator also gives startups $100,000 in seed funding, and during the second phase of the program brings them to San Francisco to conduct pitches and sales. While reporting on Shenzhen’s hardware ecosystem in 2014, I visited Hax and observed a mash-up of international entrepreneurs prototyping new products: a Brooklyn-based startup building pipetting robots for synthetic biology, and a Vienna-based startup building simple robotics kits for kids.
For entrepreneurs looking to prototype, build, and tweak a hardware product, there’s no better place than Shenzhen. Flush with factories that spent decades churning out all kinds of goods for export, the city has developed its own unique manufacturing ecosystem. The extreme density of skilled industrial engineers, dizzying electronics markets, and flexible manufacturing systems dramatically shorten product cycles. Entrepreneurs say that a week working in Shenzhen is equivalent to a month in the United States. As a result, the city—what many now dub “the Silicon Valley of Hardware”—has become a magnet for entrepreneurs and engineers who make physical products.
And therein lies the real downside in a broad-brush condemnation of “Chinese VCs.” The danger is not that Chinese VC funding in Silicon Valley will dry up. Chinese money has made quite a splash in recent years, but Chinese capital controls have already substantially reduced those flows, a downward trend that has been accelerated by the recent political backlash in the United States. Even if Chinese VC funding disappeared entirely, it would not be a devastating blow to overall funding for US innovation.
Instead, it’s the isolation from some of China’s own technological ecosystems that may harm US companies. Yes, Silicon Valley has led the way in an era of purely digital products: social media platforms and phone apps that can be “manufactured” in a college dorm room. But that near-monopoly on access to the cutting-edge is likely to shift with the emergence of next-generation technologies that need to be physically built: intelligent robotics, autonomous drones, and other real-world applications of AI. It’s no accident that the company dominating the drone industry is based in Shenzhen. Savvy Chinese investors (and accelerators like Hax) often provide a crucial gateway for small American startups to access Shenzhen’s wealth of resources (talent, ideas, and manufacturing facilities), one that could easily be sealed off.
Even for startups with no hardware element, Chinese VC funding is often taken on for specific reasons: market access in China, partnerships with a Chinese firm like Alibaba, or insights from the VC on innovative business models in China. The most promising US startups can be picky about who they take money from, and the above value-adds are some of the few reasons why a sought-after startup would opt for Chinese money over more prestigious American VC firms.
None of this is to say that oversight or restrictions on Chinese VC investments in the United States aren’t warranted. The authors of the DIUx report lay out how key dual-use technologies are now emerging out of the private sector, rather than government labs. Meanwhile, Yunan Zhang at The Information has highlighted the arrival of government-backed Chinese VCs in Silicon Valley. These state-backed investors, however, still constitute one of the least active segments of Chinese VCs—I’m told by sources in the Valley that with no track record or contacts, these government-backed VCs simply can’t get into any good deals.
Given trends in emerging technologies and the Chinese government’s clear appetite for technology transfer, the United States will need to be more vigilant and create new mechanisms for protecting nascent innovations. But when creating those mechanisms, policymakers should be clear on the goals, the targets, and the costs of imposing restrictive measures. Are we worried about any investor with operations in China? Or is it based on the nationality of the firm’s founder? Or the identity of the limited partners whose money the VCs invest?
Depending on how one answers those questions, the resulting restrictions will likely differ based on the intended target. Hax may operate in China, but it was founded and is largely run by European citizens. Sinovation is based in China, but was founded by a Taiwanese national, and the limited partners for its US funds are primarily American organizations. Other “Chinese” VCs in the Valley are composed of a patchwork of Chinese, American, and international founders and funders.
Washington may decide that the country is better off blocking all VC investments with any hint of a China connection. Most analysts describe this as a brewing “tech cold war,” and perhaps these distinctions are too nuanced for times of “war.” But if the United States make that choice, it should do so with a clear understanding of the players and the stakes involved and a sound appraisal of the costs and benefits to America’s own innovation ecosystem.