Original · Unique Research · 2026-04-16
Editor's note: The first-person report and its judgments belong to the original Chinese author. This English rendition retains the opening essay, all four thematic sections, and the complete three-question Q&A. All valuation figures, market data, company claims, and investment theses are source or interviewee claims, not independently audited findings. Company and personal names are transliterated where official English forms remain unverified. The "China asset is coming back" framing, HKEX figures, model valuations, and all robotics/Agent infrastructure projections are preserved as stated by the source.
Unique Awards
After the Valuation Surge: How USD Investors See China's AI Breakout Map
When the market reprices Chinese tech assets, what is truly worth asking is not just "who went up," but in the next round of global competition, what exactly does China's AI have to win with.
"
China asset is coming back. Driven by two turbines, Chinese tech assets are moving toward the center of the stage.
In March 2022, when most people were still figuring out what a large model was, someone was already excited enough to rush-register the WeChat ID "e/acc" and change their name to Louis.ai.
He is Liang Sicong, who prefers to be called Louis. As an AI-Native Gen Z USD investor, when a crowd of Ivy League graduates were sharpening their heads to squeeze into Silicon Valley, his first career stop was set in Jakarta and Bangalore.
Researching milk tea shops on the streets of Jakarta, examining logistics structures in Southeast Asia—you can hardly imagine this is the daily routine of an investor who watches cutting-edge AI technology. But this precisely constitutes his extremely hardcore foundation.
During the past two years of fluctuating market sentiment, he said something to me with great impact: "China asset is coming back. Driven by two turbines, Chinese tech assets are moving toward the center of the stage."
Not going global is seeking death; blindly going global means dying faster. In this wave rolling toward the globe, what kind of products can survive? What kind of capital can endure at the table? This may be a deep dive into the "disenchantment and reshaping" of going-global investment.
"The Two Turbines": Pricing Anchor Reconstruction and the Robotics Inflection Point
When discussing "why Chinese assets," Louis gave a very counterintuitive perspective.
In the past, when we talked about going global, it was mostly "supply spillover"—domestic sales were slow, things were cheap, so ship them overseas. But at the 2026 juncture, the logic has changed.
The first signal is the emergence of a pricing anchor.
Louis did a calculation: "From the second half of 2025 to early 2026, a batch of underlying model companiesconcentrated listed, with valuations pulled to around 300 billion. This directly injected enormous liquidity into Chinese AI assets." In the past, except for NVIDIA and Google, Chinese models had no benchmark ruler in the eyes of overseas assets. "Now that thisanchor has been planted, the primary market—like Kimi's valuation—has correspondingly been pulled above 100 billion. International investors will naturally judge the water temperature."
Data does not lie. In 2025, HKEX's total market cap grew 34% year-over-year, leaving those established Western exchanges behind. Foreign ownership ratios have also quietly climbed from the bottom to nearly 50%.
The second signal is the "real swords and spears" underlying position.
After OpenClaw came out, the inner circle exploded. But what Louis saw was the result of "training" behind it: "Several of China's open-source models, like Step 3.5 Flash, have consumption volumes directlyoutweighed US models. Apart from China and the US, no other market already has ten-thousand-card clusters, plenty of senior engineers, and extremely rich scenarios—they no longer have a seat at the table."
But this is not yet the most ferocious. What truly excites him is robotics assets.
"Do you know why it must be China? Because machines are crossing the valley of death from demo to productivity." Louis himself invested in ingenious, a company making dexterous hands. In the past, industrial mechanical equipment relied on manually hard-coded rules and debugging. Now AI has become the true brain. And from the supply chain, low-cost mass-production capability, and engineer density, no single country cancaught thisoverwhelming dividend.
"This is not selling a component; this is actually an independent terminal." Using hands to do work—this is the most thorough generational shift in the manufacturing paradigm.
Put Away the Wolf-Warrior Mindset, Let's Talk About Ecosystem Control
As soon as globalization comes up, everyone likes to compete on scale. Whoever ships more is awesome.
Louis waved his hand: "Scale is a result, never a cause."
Recently hespecifically went to Kyoto to study Nintendo. He admitted he had been wrong before, thinking it was just a lousy company making game consoles. "Not at all. It is a company centered on IP operations, externallyoutput ecosystem value." Mario has been hot for 40 years, and even more terrifyingly, itsmerchandise monetization can account for more than half of total revenue. At its peak, Nintendo was makingcrazy money from Europe and America byoutput culture.
"Good business futures all create new pathways," he said.
How to create? Take Pop Mart as an example—do you know Tycoco?
Most people in China have absolutely never heard of this thing. It is a blind box under Labubu specifically designed for the Mexican market. The skeleton image is considered quite unlucky in Chinese eyes, but hits the sweet spot of Mexican local culture. Finding the local context, redoing design, supply chain, and marketing—this is ecosystem position.
3D printing is the same. Hardware is just a shell; if you can use AI and 3D printing at home to generate a unique secondary-creation IP model, with the cloud automatically settling copyright fees, and material consumption being another monetization… isn't this the Bandai of the next era, the next Disney?
"Radar and Low Gear": Being Grounded Is the Best Moat
Since the direction is here, who can make it happen?
This is also the cruelest part: many tech大牛 go overseas and directly suffer from culture shock. Louis has seen too many projects that are flawless on PPT but die on foreign streets. He summarized that founders who can win global battles must have "nian" (intent/conviction), and must understand "division of labor."
What is "nian"? It is near-intuitive cross-market insight. For example, pool robots—Chinese founding teams have never lived in a house with a large pool their whole lives; how can you体会 the pain of sweeping dead leaves? Unless you personally move in locally and do first-hand research.
When Louis went to Indonesia to look at cars, he found that local carport roofs all have air-conditioning vents, because it is so hot there that cold air sinks, and the ventilation design must be redone.
"And the two-wheeled electric vehicles that sell特别 well in China—they go cold as soon as they cross the sea." In Jakarta he grabbed a rider and asked what they feared most: this place floods every day, will the electric bike leak electricity and electrocute people if submerged? "We think everyone is competing on range and price; what they care about is saving their lives. If you don't go listen on the ground, you'll never figure it out in this lifetime."
Do a good job of cross-cultural division of labor in the team, tolerate local employees' different habits, first use the high-flying "radar" to look globally for wind directions, then immediately shift into heavy-operations "low gear" to chew through the mud when landing. This is real globalization.
Agent Infrastructure Going Global: The Opportunity to Smash Old Idols
Near the end of the interview, we circled back to AI Agents.
On how AI will restructure enterprises, Louis's view is very direct: this is a process of gradually clearing out the incumbents of the previous generation of software.
"For example, Claude recently cracked COBOL. This fossil-level language was a cash cow held in the hands of big companies like IBM. Now the barrier has been smashed flat by AI, and even tiny consulting firms can get in and grab a piece." When monopolies are cut open, new standards are established.
Louis has recently also set his sights on a company making Agent identity protocols. In the future, thousands of Agents with independent personalities will be flying everywhere—who will prove "it is it"? This requires an underlying system for authentication, ID issuance, and multi-party communication.
The email protocols and H5 standards of the past were all rules set by the West. In this naturally global new track of Agents, "Chinese companies有可能 pull off the job of defining standards. Visionary founders will wait for the wind."
He finally gave a sentence to all entrepreneurs who want to be in this field:
"Born by Chinese, born for global."
See through the cracks in business with a Chinese perspective, and the rest is to bravely sail to every corner.
Selected Q&A
Q1: Capital pursues rapid scaling, but true globalization requires slowing down for local deep cultivation—how do you view this tradeoff?
Louis Liang Sicong: This is very interestingly put. Food delivery is a heavy-operations matter, and simultaneously must accommodate many local characteristics. For example, overseas urban design is different from China's high density—it is very dispersed, which means delivery times get longer; at the same time, labor is not as abundant as in China. As the saying goes, "haste makes waste." What we pursue is not blanket expansion. In practice, I will first use a certain type of store as a pilot, calculate the UE (unit economics), look at street stores and university stores separately, and select the ones with the best ROI and payback period. Always keep an eye on cash flow—good CEOs must know how to calculate this.
Q2: For globalizing enterprises, what kind of problems deserve capital to give them more patience?
Louis Liang Sicong: Focus on what you are doing now that adds points for the long term, things that are hard first and easy later. For example, infrastructure in the Agent era—who issues identities, how to manage communication. The previous generation of software relied on APIs; now if OpenClaw saves Agents with "personalities," this is like the jewelry industry needing a third-party company like IGI to do certification. This complex underlying protocol takes time to explode, but it can form an extremely high compounding moat. This is worth our patience.
Q3: In the next 3 years, what kind of capital do you think will dominate the globalization track?
Louis Liang Sicong: I firmly believe a batch of "Global Native" investment institutions will emerge. Previously when we talked about going global, it was pushing products outward from China—a supply-side perspective. But in the future, although institutions may be based in China, they will definitely use a global allocation perspective to重新 find opportunities overseas. The model will also be very "Eastern thinking"—not only giving money and venue, but also deeply binding with startup teams using China's operational experience to fight together. This will be a new school of overseas investment.