AI in China 🇨🇳

 More

FILE PHOTO: People visit a metaverse booth during World Artificial Intelligence Conference, following the coronavirus disease (COVID-19) outbreak, in Shanghai, China, September 1, 2022. REUTERS/Aly SongREUTERS

By Alexandra Alper

WASHINGTON (Reuters) - U.S. investors including the investment arms of Intel Corp and Qualcomm Inc accounted for nearly a fifth of investments in Chinese artificial intelligence companies from 2015 to 2021, a report showed on Wednesday.

The document, released by CSET, a tech policy group at Georgetown University, comes amid growing scrutiny of U.S. investments in AI, Quantum and semiconductors, as the Biden administration prepares to unveil new restrictions on U.S. funding of Chinese tech companies.

According to the report, 167 U.S. investors took part in 401 transactions, or roughly 17% of the investments into Chinese AI companies in the period.

Those transactions represented a total $40.2 billion in investment, or 37% of the total raised by Chinese AI companies in the 6-year period. It was not clear from the report, which pulled information from data provider Crunchbase, what percentage of the funding came from the U.S. firms.

Qualcomm Ventures and Intel Capital were involved in 13 and 11 investments in Chinese AI companies respectively, outpaced by GGV Capital which led U.S. firms with 43 total investments in the sector, the data showed.

The Biden administration is expected to unveil an executive order this year curbing some U.S. investments in sensitive Chinese tech industries, as hawks in Washington blame American investors for transferring capital and valuable know-how to Chinese tech companies that could help advance Beijing's military capabilities.

According to the report, U.S. investor GSR Ventures invested alongside China's IFlytek Co Ltd in a Chinese AI company after the speech recognition firm was added to a trade blacklist. Silicon Valley Bank and Wanxiang American Healthcare investments group made investments in Chinese AI firms alongside China's Sensetime before the powerhouse in facial recognition technology was added to the same trade blacklist.

Both companies were added to the blacklist, which effectively bars them from receiving U.S. tech exports, in 2019 for alleged human rights violations related to the repression of Uighur Muslims.

Some of the largest investments include Goldman Sachs' solo investment in 1KMXC, an AI-enabled robotics company, as well as an investment by three U.S.-based VC firms in Geek+, an autonomous mobile robot company, the report showed.

Only one Chinese AI company that received funding from U.S. investors is involved in developing AI applications for military or public safety uses, according to CSET.

文件照片：2022 年 9 月 1 日，在中国上海，冠状病毒病 (COVID-19) 爆发后，人们在世界人工智能大会期间参观虚拟展位。路透社/Aly SongREUTERS

 亚历山德拉·阿尔珀

 华盛顿（路透社）——周三的一份报告显示，2015 年至 2021 年，包括英特尔公司和高通公司投资部门在内的美国投资者占中国人工智能公司投资的近五分之一。

 乔治城大学科技政策小组 CSET 发布这份文件之际，美国对人工智能、量子和半导体的投资受到越来越多的审查，拜登政府准备公布对美国资助中国科技公司的新限制。

 报告显示，167 名美国投资者参与了 401 笔交易，约占同期中国人工智能公司投资的 17%。

 这些交易的投资总额为 402 亿美元，占中国人工智能公司 6 年期间融资总额的 37%。 这份从数据提供商 Crunchbase 获取信息的报告并不清楚，有多少比例的资金来自美国公司。

 数据显示，高通创投和英特尔投资分别参与了 13 笔和 11 笔对中国人工智能公司的投资，超过了纪源资本 (GGV Capital)，后者在该领域的投资总额为 43 笔，领先美国公司。

 预计拜登政府今年将公布一项行政命令，限制美国对敏感的中国科技行业的部分投资，华盛顿鹰派指责美国投资者向中国科技公司转移资本和宝贵的专有技术，这些公司可能有助于提高北京的军事能力。

 报道称，在一家中国人工智能公司被列入贸易黑名单后，美国投资者金沙江创投与中国科大讯飞有限公司共同投资了这家中国人工智能公司。 在面部识别技术巨头被列入同一贸易黑名单之前，硅谷银行和万向美国医疗保健投资集团与中国商汤科技一起投资了中国人工智能公司。

 2019 年，这两家公司因涉嫌与镇压维吾尔族穆斯林有关的侵犯人权行为而被列入黑名单，实际上禁止它们接收美国技术出口。

 报告显示，一些最大的投资包括高盛对人工智能机器人公司 1KMXC 的单独投资，以及三家美国风险投资公司对自主移动机器人公司 Geek+ 的投资。

 据 CSET 称，只有一家获得美国投资者资助的中国人工智能公司参与开发用于军事或公共安全用途的人工智能应用。

The Harvard Gazette

Why China has edge on AI, what ancient emperors tell us about Xi Jinping.

In a recent U.S. government ranking of companies producing the most accurate facial recognition technology, the top five were all Chinese companies. China is currently installing facial recognition equipment across entrances to many of its subway stations.

Recent event examines what social sciences can tell us about rising economic, geopolitical power.

Christy DeSmith, Harvard Staff Writer

March 16, 2023

Dictatorships and authoritarian regimes tend to trail more democratic and inclusive nations in fostering cutting-edge, innovative technologies, such as robotics and clean energy.

Artificial intelligence may prove an exception, at least in China, owing to dovetailing interests.

Harvard Economics Professor David Yang  spoke to the outsized success of China’s AI sector at a recent dean’s symposium on insights gleaned from the social sciences about the ascendant global power. As evidence, he cited a recent U.S. government ranking of companies producing the most accurate facial recognition technology. The top five were all Chinese companies.

“Autocratic governments would like to be able to predict the whereabouts, thoughts, and behaviors of citizens,” Yang said. “And AI is fundamentally a technology for prediction.” This creates an alignment of purpose between AI technology and autocratic rulers, he argued.

Because AI heavily depends on data, and autocratic regimes are known to collect vast troves of it, this advantages companies with Chinese government contracts, which can turn around and use state data to bolster commercial projects, he added.

Yang’s research shows China exporting huge amounts of AI technology, dwarfing its contributions in other frontier technology sectors. Yang also demonstrated that autocratic regimes around the world have a particular interest in AI. “AI quite startlingly is the only sector out of the 16 frontier technologies where there’s disproportionately more buyers that are weak democracies and autocracies.”

“Autocratic governments would like to be able to predict the whereabouts, thoughts, and behaviors of citizens. And AI is fundamentally a technology for prediction.”— David Yang, associate professor of economics

And just when are these countries most likely to buy the technology from China? Yang ended his symposium talk by mapping the uptick in purchases that follow political unrest and protest events. “To the extent that technology is exported,” Yang concluded, “it could generate a spreading of similar autocratic regimes to the rest of the world.”

Hosting Yang’s presentation was Lawrence D. Bobo, Dean of Social Science and W.E.B. Du Bois Professor of the Social Sciences. Launched in 2021, these virtual symposia gather scholars from across the division to trade research and thinking on topics of broad interest. “China in Focus: New Social Science Approaches,” which was held earlier this month, was moderated by Mark C. Elliott, the Mark Schwartz Professor of Chinese and Inner Asian History and vice provost for international affairs.

More bold predictions came from Professor of Government Yuhua Wang, whose current research relies not on contemporary economic data but ancient indicators.

Drawing from his recent book “The Rise and Fall of Imperial China: The Social Origins of State Development,” Wang shared a chart of emperor assassinations across 2,000 years of Imperial China. Gathering this data meant analyzing the biographies of nearly 400 Chinese emperors, from the Qin Dynasty to the Qing Dynasty. Turns out, about a quarter were assassinated by members of their own government and most likely during economically strong governments, hitting their peak circa 900 A.D. during the late Tang Dynasty.

“Why do we see this contradiction between the strength of the ruler and the strength of the government?” Wang asked. “Chinese rulers — historically but also contemporarily — face a tradeoff that I call the Sovereign’s Dilemma.” That is, a coherent set of government elites is capable of strengthening the state but equally capable of overthrowing the boss.

On the other hand, fragmented elites spell longevity for rulers and decline for states. This is the very dynamic Wang sees playing out today under Chinese President Xi Jinping, whose anti-corruption campaign threatens government insiders with investigation and arrest.

As evidence of splintering elites, Wang cited the sudden pivot from China’s zero-COVID policy and the recent appearance of spy balloons in the U.S. airspace. “It’s very clear the people sending balloons, maybe in the military, were not talking to the Foreign Ministry who were about to welcome [U.S. Secretary of State Antony] Blinken” for an official visit, Wang said.

“What happens is probably a very dramatic but also gradual decline of the capacity of the Chinese state.”

Also featured at the two-hour symposium was Victor Seow, assistant professor of the history of science, who covered 100 years of intensive energy extraction under multiple regimes in the country’s northeast. Ya-Wen Lei, associate professor of sociology, unpacked the human costs of China’s speedy transition from labor-intensive manufacturing to a science- and technology-driven economy.

Professors like these put the Division of Social Science in a strong position, Bobo noted at the end. “Harvard will be at the forefront of China scholarship for years to come.”

■AMD stock spikes after company launches AI chip to rival Nvidia.

Lisa Su displays an AMD Instinct MI300 chip as she delivers a keynote address at CES 2023 in Las Vegas, Nevada, on Jan. 4, 2023.

David Becker | Getty Images

AMD

 shares rose 9.9% Thursday to close at $128.37, marking the stock’s best day since May and the highest close since June. The surge comes a day after it launched new artificial intelligence chips that will compete against Nvidia

 to power AI applications.

On Wednesday, AMD CEO Lisa Su discussed the previously announced Instinct MI300X, a large graphics processor designed for AI-oriented servers, and said Microsoft

 and Meta

 had committed to using the chip.

Nvidia has dominated the AI chip market for the past year, but cloud providers and technology companies have been searching for an alternative to save costs and provide flexibility.

Thursday’s rise in AMD shares suggests investors believe the chipmaker can take a chunk of the AI chip market from Nvidia, although the company projects only $2 billion in AI GPU sales in 2024 — lower than market expectations for Nvidia AI revenue. Wall Street expects Nvidia to post more than $16 billion in data center sales in the current quarter alone, although that metric includes other chips besides AI GPUs.

AMD’s new high-end chip starts shipping in significant quantities next year.

“We believe that today’s event highlighted how AMD remains extremely well positioned to take advantage of the rapidly expanding AI TAM, as they continue to stack up customer partnerships and roll out products with impressive (and extremely competitive) performance metrics,” Deutsche Bank analyst Ross Seymore wrote in a note Thursday.

Citi analysts estimated in a note Thursday that AMD could end up with about 10% of the total AI chip market.

Su said at the launch event Wednesday that the company believes the total market for AI chips could climb to $400 billion over the next four years, twice as high as the company previously believed. Su suggested to reporters that AMD doesn’t need to beat Nvidia to do well in the market for AI chips because it will be so large.

“I think it’s clear to say that Nvidia has to be the vast majority of that right now,” Su told reporters Wednesday, referring to the AI chip market. “We believe it could be $400 billion-plus in 2027. And we could get a nice piece of that.“

2023 年 1 月 4 日，Lisa Su 在内华达州拉斯维加斯举行的 CES 2023 上发表主题演讲时展示了 AMD Instinct MI300 芯片。

 大卫·贝克尔 | 盖蒂图片社

 AMD 股价周四上涨 9.9%，收于 128.37 美元，创下该股自 5 月份以来的最佳单日表现，也是自 6 月份以来的最高收盘价。 就在一天前，该公司推出了新的人工智能芯片，该芯片将与英伟达竞争，为人工智能应用提供动力。

 周三，AMD 首席执行官苏姿丰 (Lisa Su) 讨论了之前发布的 Instinct MI300X，这是一款专为面向 AI 的服务器设计的大型图形处理器，并表示微软和 Meta 已承诺使用该芯片。

 过去一年，英伟达一直主导着人工智能芯片市场，但云提供商和科技公司一直在寻找替代方案来节省成本并提供灵活性。

 AMD 股价周四上涨，表明投资者相信这家芯片制造商可以从 Nvidia 手中夺取 AI 芯片市场的大部分份额，尽管该公司预计 2024 年 AI GPU 销售额仅为 20 亿美元，低于市场对 Nvidia AI 收入的预期。 华尔街预计 Nvidia 仅在本季度的数据中心销售额就将超过 160 亿美元，尽管该指标还包括 AI GPU 之外的其他芯片。

 AMD 的新型高端芯片将于明年开始大量发货。

 德意志银行分析师 Ross 表示：“我们相信，今天的活动突显了 AMD 在不断积累客户合作伙伴关系并推出具有令人印象深刻（且极具竞争力）性能指标的产品的过程中，仍然处于非常有利的地位，能够充分利用快速扩张的 AI TAM。” 西莫尔周四在一份报告中写道。

 花旗分析师在周四的一份报告中估计，AMD 最终可能会占据整个 AI 芯片市场约 10% 的份额。

 苏在周三的发布会上表示，该公司相信人工智能芯片的总市场将在未来四年内攀升至 4000 亿美元，是该公司此前认为的两倍。 苏向记者表示，AMD不需要击败Nvidia才能在AI芯片市场取得好成绩，因为这个市场规模太大了。

 “我认为很明显，英伟达现在必须占据其中的绝大多数，”苏周三在谈到人工智能芯片市场时告诉记者。 “我们相信到 2027 年，这一数字可能会超过 4000 亿美元。我们可以从中分得一杯羹。”