In the Dragon’s Lair: My Week in a Chinese Venture Capital Firm

It’s likely not often that an Air Force intelligence officer spends a week inside a Beijing-based Venture Capital firm with support from all sides. I had this remarkable opportunity as a student in the Massachusetts Institute of Technology's Executive MBA Program, where students have the option to work with an international client in their final semester. Since I visited China as an MIT student, this is probably a good time to say that nothing in this article represents the views of the Department of Defense.

It’s not hard to find credible articles about the Chinese threat to U.S. national security, from investing in Silicon Valley startups to taking the international lead in 5G deployment to even Chinese government's use of U.S. commercial space assets. Is this just a series of smart investments by a rising power, or is this evidence we're inching toward Thucydides’s Trap? I wanted to find out, and the only way to do that was to see it for myself.

Many discussions with security-conscious colleagues suggested that, upon landing in Beijing, I should expect anything from a beating with a printed copy of my security clearance forms, bags of cash, or bugs on all of my electronic devices thanks to the Chinese state. I’m very happy to say that none of that occurred. Our hosts took great care of us, and the only bug I experienced was in the food poisoning that hit me 5 minutes before take-off back to the US. Pro tip: avoid open-air airport dumplings of unknown provenance. 

In addition to the food poisoning, I returned home with an appreciation of how far China has come -- and how masterfully they can exploit the West's investment in technology and institutions. Here were my biggest surprises:

• The average Chinese Government contribution to a single fund in one VC (40% of $150MM, or $60MM) is more than double the entire Department of Defense’s budget for investment in commercial innovation through the Defense Innovation Unit ($29MM).
• Beijing tech startups outsource labor to Western nations (e.g. Germany) because software developers are less expensive and more loyal in the West than in Beijing. They also open “research centers” in the West as a brilliant move to more easily poach AI talent.
• Chinese companies have access to an incomprehensible amount of data and have the freedom to use this data to train algorithms in ways that are not possible in the West. For example, one company has 40 petabytes (40,000,000 Gigabytes) of videos of kindergarteners being used to train computer vision algorithms.
• China has the near-term potential to lead in algorithm development and exportation; it’s feasible to think that, in the not too distant future, the West will purchase black-box Chinese algorithms that simply work better than anything else due to the volume and variety of data used to train them.
• China understands the importance of winning the 5G race and is helping startups prepare for it. Huawei actively shares geospatial and other data they can use to make their apps/services 5G-ready. And there are 5G towers all over Beijing. Today.

About Our Host

Our host was GoldTech, a Beijing-based Venture Capital firm that manages assets worth more than 10 billion Yuan ($1.5B USD). They invest in more than 100 companies across China in various high-tech sectors such as Artificial Intelligence (AI), Big Data, Blockchain, Fintech, and energy. They understand AI at a fundamental level and even have a multi-million-dollar AI fund to make tailored international investments. They also have a working knowledge of English -- as evidenced by their prior work and master’s degrees from the United States. They could easily hold their own in a U.S. investment firm.

They also lease space to startups in their 1 million square foot Aobei Science and Technology Park (pictured above). By combining investment in seed-stage companies with real estate, they hope to create virtuous cycles that creates more value than the efforts would independently. It seems to be working - according to their website, eight of these companies are listed on stock exchanges, and as many as sixteen are close to an IPO. 

Chinese Investment Abroad

The national security concerns of Chinese investment are legitimate: with an ownership stake and potentially seats on the board, investors can influence the operations of a company in ways counter to U.S. interests. But more than that, Chinese VCs want the best returns for their limited partners, and the U.S. is a great place for returns on tech investment. Making better investments abroad and understanding Western problems was one of the main reasons they wanted an MIT team to advise them. They also have another connection to MIT - as a limited partner in the Boston Angel Club (BAC), they are able to target startups coming out of MIT and Harvard through a U.S. fund. It’s perfectly legal, and it provides Chinese investors a way to make early investments in high-reward emerging tech areas like AI, robotics, and data. But there are always strings attached.

State Support and Capital 

Our host VC’s goal is to place about a third of each fund in overseas investments. On average, their funds are approximately 1 billion Yuan (~$150MM USD), and approximately 40% of funds come from a combination of local and national government. For example, they told me that Beijing invests approximately $10 billion annually in funds like these. VCs are expected to invest in state priorities for their government investment, much as Corporate VCs (CVCs) invest in line with company priorities. For a single $150MM fund, approximately $60MM comes from the Chinese state.

Let’s compare this to how the U.S. government directs investment in private companies. My organization, the Defense Innovation Unit (DIU), received $29MM this year in what’s known as RDT&E funding – this is the working capital to deploy non-dilutive contracts to the best emerging tech companies. The entire Department of Defense’s investment through DIU is less than half of the Chinese government’s contribution to a single fund in one VC. There are many, many VCs in China. From a dollars and deals perspective, it’s not that China and the U.S. aren’t in the same ballpark – I don’t think we’re playing the same game.

Meeting with Chinese Startups

We met with five portfolio companies while in Beijing, and it was clear that they view the West more as a business opportunity than as the enemy in a great-power competition. I don’t remember any mention of the “trade war,” but there were some jokes directed at President Trump about his struggle to build a Great Wall of his own.

All of the companies claimed to use some aspect of artificial intelligence, and most had an eye toward how they will adapt to the roll-out of 5G. Most had an eye toward Western expansion – either by hiring Western software developers because they were cheaper and more loyal than locals or by opening “research centers” in the West to poach AI talent.

These Chinese startups have some really interesting things going on:

• One company has a trial with a police department to integrate geospatial data, building plans, local utilities and police systems. A terrifying and dystopian example was that, if an apartment’s power was off for an extended period of time, the police would be alerted to do a health and wellness check. This almost perfectly tracks this MIT Tech Review article aptly entitled, “Who needs democracy when you have data?”
• One company can precisely quantify the value of social media influencers through AI. Sentiment analysis shows how enthusiastic an influencer is about a sponsor product or company. Computer vision shows how prominent logos are in a video. Language processing helps companies know how to market based on what is popular with influencers. And of course the government is a customer!
• One company makes educational robots that constantly record kindergartners and upload it into the cloud. Algorithms are being used give insight into their emotion, concentration through eye focus, and activity. With 40 Petabytes of video of kindergarteners, they know they have an unprecedented amount of training data to learn more about how children learn and how to market to them in the future. 
• One company aggregates data for users across social media platforms and has more unique users than the population of the United States. 

Observations in Beijing

Cameras. Cameras everywhere. Every entrance, every public area, every hallway. Many of the cameras were the exact same model, suggesting that there is larger coordination of the streams. You constantly get the feeling you're being watched, likely because you are. Here's a shot from near Tiananmen Square:

The Chinese Government doesn’t shy from using their surveillance as an overt tool of oppression in the western part of the country, but there is nothing preventing from using this data in other ways. Of course, this is a dream for training computer vision algorithms. It’s feasible to think that the government can track you down within seconds - whether it’s through a facial match, license plate match, or in one of the many situations where your passport / national ID card has to be scanned. I much prefer the US, where only Facebook can track me down within seconds. 

I got to see a subset of this in action: one of the VC's portfolio companies uses facial recognition for building access. In a twist of irony, they said it doesn't work well on Westerners. It appears that AI bias isn't limited to the West!

I was amazed, and somewhat disturbed, at how well-developed Beijing is as a major city. Traffic is a nightmare, but they are investing a tremendous amount of money in infrastructure and new buildings:

Pollution is also horrific. These photos were taken from the same location, 2 days apart:

The Beijing subway is a different story. It's actually cleaner than the Washington DC Metro and does a better job of integrating technology. It somehow felt easier to navigate the Beijing subway without speaking any Chinese than navigating the DC Metro as an English speaker. Here's a photo from one of the stations:

Final Thoughts

Even though I wasn't exposed to it during my short time in Beijing, the real concern is the investment into dual-use technology to benefit the Chinese military. China is investing hundreds of billions of dollars in semiconductors, AI, and other dual-use tech with no clear delineation between commercial and military use. Capital is being deployed thoughtfully as part of a systemic plan, rather than a fiscal year at a time as we tend to do in the U.S. Built on systemic intellectual property theft, combined with a highly entrepreneurial spirit, China is able to take what the West has developed over generations and use it to their advantage.  

We should all be concerned that Congress’ conviction for commercial defense innovation extends to less than half of what the Chinese government puts into a single VC fund. This is not the first time that the United States as faced what some may see as an existential challenge from overseas. We rose to the challenge during the Cold War through the first and second offset strategies, and it’s time to regain that sense of urgency. The U.S. Government has to do a better job stimulating investment into dual-use technology critical to future military needs. These areas are underserved by U.S. investors, but that's not the case with Chinese investors supported by the Chinese government.

We can do better. And after a week in a Chinese VC firm, I realize that we have to do better.