China's Hermitage Capital bets big on AI as it sees tech shaping the future
The euphoria surrounding AI will define this decade, according to Sean Xiang Yuqiu, founder of
Hermitage Capital . 'It's just like 1995, when the internet era took off and ran for 20 years, giving rise to legendary companies like Yahoo, Amazon and Google,' he said in an interview with the Post recently.
Xiang, who has been featured in the latest edition of Fortune China magazine's 40 Chinese business elites under 40 released on Tuesday, said Hermitage would invest at least US$500 million in the AI and robotic sectors over the next three years.
With a decade of experience at Wall Street investment bank JPMorgan Chase before founding Hermitage Capital in 2017, Xiang said the company's investments focused on building a global AI ecosystem, ranging from hardware like graphics processing units to large language models (LLMs) and intelligent robotics.
03:16
China shows off latest AI innovations at international conference in Shanghai
China shows off latest AI innovations at international conference in Shanghai
He drew a parallel to humanoid robot making to illustrate his investment strategy. 'You need a brain to process information, eyes to observe the world and blood as your power source,' he said. 'It must also sense the external world and respond to it effectively. We've invested across all these areas, from foundational AI models to robot chips to solid-state batteries and home robots that free people from tedious chores.'
Computer vision led the first wave of AI's evolution, which was now largely mature, Xiang said. 'We are currently riding the natural processing language wave, dominated by LLMs. The foundation models' valuation will still have a five to 10 times return potential.'
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
South China Morning Post
7 minutes ago
- South China Morning Post
How does Tam Jai's new HK$18 canned zhajiang meat topping compare to the fresh stuff?
Best known for its steaming bowls of noodle soup, Hong Kong restaurant chain Tam Jai Yunnan Mixian has launched a canned version of its zhajiang topping, a fermented bean and minced meat sauce. Zhajiang is a popular topping at Tam Jai, which takes inspiration from Hong Kong's pick-and-mix 'cart noodle' shops . In operation since 1996, it has more than 240 locations, mostly in Hong Kong but also in mainland China, Singapore, Japan and Australia. It is well known for its Tam Jai jeh jehs, the waitresses who make up much of its workforce (jeh jeh means 'older sister', or auntie, in Cantonese). Sometimes called Chinese or Beijing bolognese sauce, zhajiang is traditionally made with fermented bean sauce and sliced or minced meat, such as pork or beef. It is often eaten with noodles in a dish known as zhajiangmian , which is said to have originated in mainland China's Shandong province. Zhajiang translates as 'deep-fried sauce' in Mandarin. It comes from the cooking process where, after a period of frying, the oil separates from the solid ingredients and floats to the top, giving the illusion that the sauce is being deep-fried. 02:53 Hong Kong noodle chain Tam Jai just released a canned version of its popular 'zha jiang' sauce, but Hong Kong noodle chain Tam Jai just released a canned version of its popular 'zha jiang' sauce, but Hong Kong versions of zhajiang sauce typically have added sugar and taste sweeter than their mainland counterparts.
South China Morning Post
7 minutes ago
- South China Morning Post
Meta drives first-half AI smart glasses surge, as Xiaomi emerges as dark horse
The global smart glasses market saw shipments more than double in the first half of the year, driven mostly by strong demand for Meta Platforms' glasses, along with new offerings from rising competitors like China's Xiaomi, according to an industry report. The 110 per cent year-on-year surge in shipments was fuelled by robust interest in the artificial intelligence glasses segment, which accounted for 78 per cent of total smart frame shipments, according to the report on Tuesday by market research firm Counterpoint. The remaining market consisted mainly of basic smart glasses, which mostly include smart audio glasses. Meta, one of the pioneers in commercialising AI glasses, led the market with a more than 73 per cent share. Other major players include Chinese companies like Huawei Technologies , TCL Electronics subsidiary RayNeo , and Xiaomi Visitors try on smart glasses during the third China International Supply Chain Expo in Beijing last month. Photo: Xinhua Analysts attributed the AI glasses segment's growth to the popularity of devices like the Meta Ray-Ban series and Xiaomi's newly launched AI glasses.
AllAfrica
37 minutes ago
- AllAfrica
AI vs jobs, China trade truths and dictator economics
Here's a roundup of recent radical research. It's practically conventional wisdom that AI is going to take jobs away from large numbers of humans, leaving them without anything useful to do in the economy. People are so convinced of this that they'll jump at practically any hint in the data that allows them to believe that it's happening. A little while ago I wrote a post about why both economists and popular commentators are getting way over their skis on this. Anyway, Sarah Eckhardt and Nathan Goldschlag of the Economic Innovation Group have a good new report on this, which shows that as far as we can tell, AI isn't taking jobs yet — at least, not on any measurable scale. Eckhardt and Goldschlag start with a measure of predicted AI exposure for various jobs. These measures don't tell you which jobs are going to be replaced by AI; instead, they just tell you which jobs currently involve more tasks that can probably be done by AI. These measures actually have a pretty good track record at predicting which workers will end up using AI. Basically, Eckhardt and Goldschlag find no correlation — or even a negative correlation — between that measure of AI exposure and any measure of labor market distress. For example, here's the unemployment rate of workers with varying degrees of predicted AI exposure (1 is the least exposed, 5 is the most exposed): Source: EIG There has been a recent rise in unemployment, but it's concentrated among the people who are least exposed to AI, while those who are the most exposed almost all still have jobs. The same is true when we look only at recent college graduates, who have been the focus of the most concern in the media: Source: EIG And the same is true when we look at which workers are exiting the labor force completely: Source: EIG And one more interesting finding is that the most-exposed workers are actually less likely to switch to less-exposed occupations than they were before generative AI hit the market! In other words, coders and paper-pushers are not becoming plumbers to protect themselves from AI: Source: EIG The researchers also try using alternative measures of AI exposure, and they find pretty much the same thing. In other words, AI job displacement just hasn't happened yet. It may happen in the future, but so far, every time people have jumped at a particular data point to claim it's finally happening, it has turned out to be a mirage. Bernie Sanders and his followers deeply believe that America's economy is in a prolonged state of crisis — that capitalist economic policies have steadily immiserated the American public, creating a country where regular people are economically drowning even as corporate fat cats enrich themselves. Their absolute faith in this narrative often leads them to interpret economic statistics in dubious or even ridiculous ways. The latest example of this is when Bernie Sanders posted a chart of housing versus wages: This is a pretty ridiculous chart. Why would you plot home prices on the same y-axis as weekly income? Does anyone think these two things should be even remotely close to the same size? Do we think people should be able to afford a house on a single week of income? That's ridiculous. A non-ridiculous way to present this data would be to divide home prices by weekly earnings. That would show us how many weeks a typical worker would need to work in order to afford a home. But actually, the 'median weekly earnings' number is for full-time workers only, so instead we should use median personal income, which counts everybody. Here's what that looks like: In the 80s and 90s, it took about 8 years of work to afford a home. Since then, the number has climbed to about 10 years — a significant and concerning drop in affordability, but not a catastrophic drop. A breakdown by the Economic Innovation Group shows that mortgages are about as affordable as they ever were, but down payments have gotten less affordable: Source: Ben Glasner While the drop in housing affordability over the past half century is certainly a problem, it's not the kind of crisis that Bernie paints it as. Using silly charts in service of alarmist narratives ultimately just weakens trust in your movement — or at least, it should. The three most powerful countries in the world are now all ruled by strongmen. In China, Xi Jinping has subdued all rivals, and concentrated what used to be a dispersed bureaucratic oligarchy under his own personal rule. In Russia, Putin is effectively an emperor. The US is still officially a democracy, but democratic norms and institutions are eroding rapidly, and in April a majority of Americans called Trump a 'dictator.' The question is what effect these personalistic regimes will have on the economy. China's growth over the past four decades pretty much proves that democracy isn't necessary for a strong or even dominant economy. But there's a difference between countries ruled by a single strongman, and countries ruled by a system of elite institutions that distribute power among a number of oligarchs. A new paper by Blattman, Gehlbach, and Yu shows that personalist regimes tend to experience lower economic growth than either democracies or autocracies with more distributed power. The difference isn't huge, but you can see it on a graph: Source: Blattman, Gehlbach, & Yu (2025) It's not clear which direction the causation runs here; it could be that countries with bad economies tend to turn to strongmen to save them. But Blattman et al. test for this using variables that tend to predict regime transitions, and they don't find any change. That implies that personalist regimes actually make mistakes that slow down economic growth. Xi, Putin, and Trump certainly don't exactly seem to be violating that rule of thumb. China's growth has slowed relentlessly under Xi, and his industrial policy seems to be simply driving Chinese companies into unprofitability rather than extricating the country from its economic slump. Putin's war in Ukraine is slowly crushing the life out of the Russian economy, while Trump's tariffs continue to wear down the resilient US economy. The trend toward strongmen is a bad one. Does it ever seem like modern political discourse is dominated by crazy idiots? Well, that's because it is. In a new paper entitled 'Dark personalities in the digital arena: how psychopathy and narcissism shape online political participation', Ahmed and Masood find that your intuition isn't wrong: This cross-national study investigates how psychopathy, narcissism, and fear of missing out (FoMO) influence online political participation, and how cognitive ability moderates these associations. Drawing on data from the United States and seven Asian countries, the findings reveal that individuals high in psychopathy and FoMO are consistently more likely to engage in online political activity…. Conversely, higher cognitive ability is uniformly associated with lower levels of online political participation. Notably, the relationship between psychopathy and participation is stronger among individuals with lower cognitive ability in five countries, suggesting that those with both high psychopathy and low cognitive ability are the most actively involved in online political engagement. Almost everyone blames recent political trends on their chosen enemy group, but the real culprit is social media, which has elevated the worst people in our society to positions of influence from which they were previously shut out. What force can defeat the terrible power of social media and its armies of crazy idiots? In my Fourth of July post, I expressed hope that AI algorithms could be harnessed to defeat the hordes of humanity's worst: LLMs give platforms the ability to cheaply and quickly filter content according to sentiment. Simply having an LLM downrank angry content and uprank positive content would lean against the natural tendencies of social media technology. Call it Digital Walter Cronkite. Of course, that solution would depend on the willingness of platform owners like Elon Musk to unleash algorithms in the service of moderation and reasonability. That seems a bit like wishful thinking, I admit. But there's another possibility, which is that AI itself will simply naturally drive humans off of social media, by generating infinite amounts of slop. A new paper by Campante et al. finds evidence that AI-generated images nudge news consumers toward more trustworthy human-gatekept media: We study how AI-generated misinformation affects demand for trustworthy news…Readers were randomly assigned to a treatment highlighting the challenge of distinguishing real from AI-generated images. The treatment raised concern with misinformation…and reduced trust in news… Importantly, it affected post-survey browsing behavior: daily visits to [the mainstream newspaper's] digital content rose by 2.5%…[S]ubscriber retention increased by 1.1% after five months…Results are consistent with a model where the relative value of trustworthy news sources increases with the prevalence of misinformation, which may thus boost engagement with those sources even while lowering trust in news content. A similar effect might happen with AI agents, which are already flooding social media with trash commentary. As X and other social media companies lose the battle against the bot swarms, human users may stop relying on those feeds for their window on the world. The psychopaths and attention-seekers might simply get drowned out in the automated cacophony. That would be a weird end to the age of mass social media, but honestly, it's not the worst ending I could think of. The American and Chinese economies continue to decouple. Even though Trump keeps 'pausing' his tariffs on China, China is selling less and less to America: Source: Bloomberg via Noahpion But you'll notice that China's exports to Europe and Southeast Asia are still growing strongly. This has led some commentators to claim that China is simply shipping its good through third-party countries, avoiding tariffs (or the threat of tariffs) by essentially just slapping a different 'made in' label on stuff that was actually made in China. Those claims are wrong. You can see that they're wrong by looking at the actual products that China is selling to countries in Southeast Asia (the region usually accused of transshipping Chinese goods to the US.), versus the products it used to sell to the US. The two sets of products don't match up very well, meaning that only a modest portion of Chinese trade with Southeast Asia could reflect diversion of trade from the US Gerard DiPippo did this exercise: Facing US tariffs, China's exports to the US are down—while exports to Southeast Asia are up. Is that trade diversion and potential transshipment? My estimate: at most 34% of the increased PRC exports to SE Asia in Q2 could reflect trade diverted from the United States. In fact, this is an upper bound. Many of the countries being accused of transshipping Chinese goods — Mexico, Vietnam, etc. — have their own industries as well, which export a lot to the US. Increased US imports from those countries are likely to at least partially — or perhaps mostly — be locally made goods. All this goes to show that you can't draw conclusions about decoupling just from macro data. There are a number of popular ideas out there about who marries whom. One is that rich men primarily want physically attractive wives and don't care about social status, education, and so on. Another is that power couples tend to be dual earners. In fact, both of these stereotypes are wrong. As Lyman Stone shows in a post for the Institute for Family Studies, rich men tend to marry highly educated, high-earning women who become housewives after marriage. Here are some charts: Source: Lyman Stone Source: Lyman Stone Source: Lyman Stone I don't like the use of the word 'overwhelmingly' in any of these charts, but the point is clear — rich men, who presumably have greater choice in who they marry, often tend to prefer women who are educated and high-income before marriage, but many of these women become homemakers after marriage. Call it the 'power trad' couple. This article was first published on Noah Smith's Noahpinion Substack and is republished with kind permission. Become a Noahopinion subscriber here.
