If there is such a person, he was admitted to the "Youth Class" of the University of Science and Technology of China at the age of 16, graduated with a doctorate at the age of 25, became a doctoral supervisor at the age of 31, and is worth nearly 300 billion yuan at the age of 41. He has achieved both academic and financial success. Why do you think he is? Surpass Nvidia with technology? Because your family has a strong background? Or is it based on capital market speculation?

None. He relies on the fact that in the wave of "stuck neck" supply cuts, he has created one of the very few "replaceable" AI training chips that can be mass-produced in China.

His name is Chen Tianshi, and the Cambrian he created has just rewritten the history of China’s Science and Technology Innovation Board.

On June 30, Cambrian closed at 1,595.55 yuan per share, with its total market value reaching 1,002.5 billion yuan, becoming the first company on the Science and Technology Innovation Board with a market value exceeding one trillion yuan. Chairman Chen Tianshi's net worth has also risen. With a 29.34% shareholding, his personal wealth is close to 300 billion yuan. If we focus on this year, the cumulative increase in Cambrian's stock price has reached 75%, which means that in just half a year, Chen Tianshi earned 120 billion.

Dancing together are the founders of AI upstarts such as Zhipu AI, MiniMax, and Zero One Wish. Their net worth has also surged by tens of billions in the past six months along with a new round of computing power surge and valuation reshuffle.

But among these technology companies, Cambrian is still the most special one - it is the first hard technology stock on the Science and Technology Innovation Board, and even the entire A-share market, with a total market value exceeding one trillion.

For a long time, the "trillion club" of A-shares has been reserved for traditional industries. Looking at the historical list, we will find that those who dominated the list before were always those who sell alcohol, open banks, and mine coal and refinery.

The emergence of Cambrian may mark a fundamental shift in the "aesthetics" of the capital market: the logic of wealth creation is shifting from "relying on resources, relying on background, and relying on industry monopoly" to "relying on technology, relying on knowledge, and breaking external blockades."


Putting aside the superficial narrative of "winning money" in the context of self-media, how can Cambrian support a trillion-dollar market value?

First of all, Cambrian came up with mass production chips.

In the past two years, Cambrian has successively launched two AI chips, Siyuan 590 and 690. Take the Siyuan 590, which shipped nearly 100,000 units last year, as an example. It targets the A100 launched by Nvidia in 2020. According to data from multiple evaluation agencies, its performance has reached about 80%-90% of the A100.

Although the Siyuan series is still a generation behind NVIDIA's latest Blackwell architecture, its mission is not to surpass, but to solve the existential anxiety of "no cards available" for large domestic model manufacturers.

Large-scale delivery has reversed Cambrian's eight-year history of losses. In 2025, the company achieved annual profits for the first time; by the first quarter of 2026, it made over 1 billion yuan in net profits in a single quarter, a year-on-year increase of 185%. This dramatic reversal from the "gold-eating beast" to the "money printing machine" instantly ignited the frenzy of the capital market.

Secondly, Cambrian has reaped the dividend of domestic replacement.

Due to well-known reasons, domestic large-scale model manufacturers cannot obtain NVIDIA's most advanced AI chips. However, the global large-scale model competition is in full swing. Each company is "expanding its military and preparing for war", and its desire for computing power has reached its peak.

For example, the product iteration speed of a large model company has recently shown signs of slowing down. Some people in the industry speculated that it was because "there were not enough cards" and the training phase was blocked.

On the one hand, there is the practical dilemma of "cannot buy", and on the other hand, there is the competitive pressure of "must count". Domestic large model companies can only turn to local AI chips. As one of the few domestic companies that can provide "replacement solutions", Cambrian has naturally taken on this huge wave of overflow demand.

Today, when computing power has become "digital oil", major domestic manufacturers and AI upstarts are queuing up to urge Cambrian to deliver.


However, the laws of physics and business reality will not change due to a trillion-dollar market capitalization. Compared with Nvidia, Cambrian still has at least three hurdles to overcome:

First, the product must go from "usable" to "easy to use."

Showing eye-catching parameters in the laboratory or at a press conference is called "usable". But in actual deployment, tens of thousands of chips have to run like a brain for a month without breaking the chain to be called "useful". When domestic AI chips make the leap from "single-card run-through" to "ten thousand-card cluster", the engineering stability of the cluster still faces tests. Once something goes wrong in the middle, the entire training task is forced to be interrupted. In the context of involution where large models have to be upgraded every 3-6 months, companies simply cannot afford to wait for such a time cost.

Second, the technology needs to move from “mid-range” to “high-end”.

This is the most helpless problem of the Cambrian. Due to limitations in advanced equipment, Siyuan series chips have to stay in the 7nm process for a long time. Even the latest Siyuan 690 must rely on advanced packaging technology to make up for the backwardness of the process. But under the physical limit of 7nm, there is a ceiling for the chip’s transistor density and power consumption. NVIDIA can integrate the world's best resources, using TSMC's 4nm process and SK Hynix's HBM high-bandwidth memory. This kind of ecological generation gap in the entire industry chain is not something that a hard technology company in the Cambrian era can "change its fate against the odds."

Third, the system must move from “chip” to “ecology”.

NVIDIA's real moat has never been how fast the graphics card is, but that it has the CUDA ecosystem that has dominated the world for nearly 20 years, and has firmly tied all large model companies to this "underlying operating system". For domestic companies, migrating business code from CUDA to Cambrian's BANG system means investing huge time and labor costs. Stronger than DeepSeek, in order to adapt to domestic computing power, it had to reconstruct its Wanka cluster software system from the bottom up, which even directly affected the release rhythm of the new model V4.

This hard constraint on the underlying software and physical limits is not only a problem for Cambrian, but also a common proposition for the entire Chinese technology industry.


Chen Tianshi collected RMB 120 billion in half a year, which makes people reflect on Wang Jianlin's famous saying: "It is better to be bold than Tsinghua University and Peking University."

That used to be the logic of making wealth in the real estate era - relying on courage to enclose land and relying on background to clear customs. But in the AI ​​era, "being bold" no longer works. "Tsinghua University and Peking University" have almost become the standard for AI upstarts. Knowledge and technology are rewriting the new logic of wealth creation in the capital market.

But on the other side of sudden wealth, Cambrian's trillion-dollar market value is a "vote of confidence" and an "active volcano."

A price-to-earnings ratio of more than 300 times means that the capital market has realized its expectations for the next few years in advance. The carnival of capital is fleeting. Once Cambrian product iteration and mass production delivery are slightly delayed, unsustainable high expectations will immediately backfire.

For 41-year-old Chen Tianshi, the most glorious first half of making wealth has come to an end; and the real hard-core protracted battle has just begun.