A rather paradoxical phenomenon is emerging in the primary market.
On one hand, the venture capital industry has continued to discuss "difficult fundraising and exits" over the past two years; on the other hand, in a few hot sectors like AI large models, embodied intelligence, nuclear fusion, and quantum computing, scenes of rapidly rising valuations and fierce competition for financing rounds have reappeared.
According to incomplete statistics from First Finance, in the first half of 2026 alone, the domestic embodied intelligence sector added 19 companies with valuations exceeding 10 billion yuan (approximately 100 billion RMB). Among them, valuations for companies like Galaxy General, Self-Variable, ZhiPingFang, and XingHaiTu have already surpassed 20 billion yuan. Some companies have completed the leap from a 10-billion to a 20-billion-yuan level within just six months.
Valuations for AI large models are even more astonishing. Moonshot AI completed its Series F funding round exceeding $3.5 billion in July this year, reaching a post-money valuation of $35 billion. It subsequently initiated a Pre-IPO round, with market disclosures indicating a target pre-money valuation has now risen to $50 billion, approximately 340 billion yuan. At the beginning of this year, its valuation was still around $10 billion.
Money is not flowing evenly back across the entire primary market.
What's really happening is:
The limited active capital is more aggressively crowding into a few projects that appear 'more certain'.
1
What the Primary Market Lacks Most Now,
Is Not Projects, But Exits
To understand this round of valuation increases, one must first understand the primary market's greatest pain point in recent years.
It's not a complete lack of good technology, nor a complete lack of capital.
It's exits.
The biggest difference between primary and secondary market investing is that investors in the primary market cannot sell at any time. When an institution invests in a startup today, realizing actual returns typically relies on an IPO, M&A, or secondary share sales.
As long as the exit path is unclear, even high paper valuations are difficult to convert into cash returns for the fund.
Therefore, when a company suddenly shows clear IPO expectations, its attractiveness to investment institutions changes rapidly.
Moonshot AI is a classic case.
Public reports show the company conducted multiple funding rounds within six months in 2026, with its valuation rising from around $10 billion to $35 billion, and initiating a Pre-IPO round targeting a $50 billion pre-money valuation. There were even market reports of financing subscriptions significantly exceeding the original plan, leading to early closing of the round.
Here, what capital is buying is no longer just "the growth of AI over the next decade."
It's something else:
Being closer to an exit.
If a company is expected to potentially IPO only years later, investment institutions bear prolonged uncertainty; if it has already begun share reform, listing tutoring, or Pre-IPO financing, then even at a higher valuation, some capital is willing to accept it.
This explains the seemingly contradictory choice in today's primary market:
The price is higher, yet institutions perceive the risk as potentially lower.
2
Public Companies Are 'Repricing'
Private Companies
The second source of certainty comes from the secondary market.
In the past, pricing for AI and robotics companies in the primary market often relied on revenue projections, technical teams, market size, and peer financing.
An increasingly important reference point now is listed companies in the same sector.
When a publicly traded peer achieves a high market capitalization, unlisted companies immediately gain a new valuation anchor.
This is also why the financing fervor for embodied intelligence noticeably heated up in 2026. Many leading companies are pushing forward their capitalization processes. After companies like Unitree Robotics garnered high market attention, investment institutions naturally began recalculating the potential value of their unlisted peers.
The logic is very simple.
Assume an unlisted robotics company is valued at 10 billion yuan. If the market believes a peer could reach 30 billion yuan after listing, then 10 billion yuan appears to have room to rise.
But the problem lies precisely here.
The secondary market can reprice daily, but the primary market struggles to do so.
Once the listed company falls from 30 billion back to 15 billion, the "safety cushion" for the unlisted company valued at 10 billion may vanish instantly.
Therefore, the secondary market can not only push up valuations in the primary market but also pull them down.
This means that while decacorns proliferate this year, the entire primary market is actually becoming increasingly dependent on public market sentiment.
3
"The Next Round Will Be More Expensive",
Is Becoming Another Kind of Certainty
More subtle than an IPO is that the transactions themselves are beginning to create certainty.
Assume a company's previous round valuation was 5 billion yuan.
If the next round raises only 500 million yuan, releasing very few shares, it might push the latest valuation to 10 billion yuan.
Existing shareholders thus achieve paper gains. New investors are willing to accept the 10 billion yuan valuation because they believe the next round might reach 15 or even 20 billion yuan.
As long as new capital keeps entering, this valuation system can continue to operate.
Thus, the investment logic may change.
In the past, institutions asked:
What will this company be worth in five years?
Now, some investors first ask:
Will someone come in at a higher price in the next round?
These two logics may seem similar but are fundamentally different.
The former relies on the cash flow and profits the company will ultimately generate.
The latter relies on the transactions being able to continue.
This is also where the primary market needs to be most vigilant.
Valuation increases alone do not prove increased value.
4
A 10.6 Billion Yuan Nuclear Fusion Valuation
Also Shows Capital Betting on 'Scarce Sectors'
This capital concentration is already spreading from AI and robotics to earlier-stage hard tech.
In July this year, Xinao Fusion conducted its first external financing round, with a post-money valuation reaching 10.6 billion yuan after its Pre-A round. The funds will be used for the construction of a third-generation spherical tokamak hydrogen-boron fusion device, technology iteration, and R&D team expansion.
Notably, this is still a field quite distant from large-scale commercialization.
Since 2026, at least 12 domestic private fusion companies have announced new financing rounds in China; from 2022 onward, these companies have had about 24 publicly disclosed financings, with 14 of them concentrated in the first seven months of this year.
Why is capital willing to assign a 10-billion-yuan valuation when the business model is not yet fully proven?
The answer, besides technological prospects, is scarcity.
When policy, industry trends, and capital consensus simultaneously focus on a sector, and the number of companies with mature teams and technical积累 is limited, capital will proactively pay a scarcity premium.
But scarcity does not equate to ultimate success.
Nuclear fusion, quantum computing, and general-purpose robots may all have trillion-yuan-level imaginations, yet they may also require years or even over a decade to complete genuine commercial validation.
5
The Primary Market Is Shifting from
"Technology Risk" to "Valuation Risk"
This also explains an important distinction between this wave of decacorn fever and the previous internet investment wave.
Today, capital does not know which embodied intelligence path will certainly win, nor which large model will ultimately establish a stable business model.
Technological uncertainty has not disappeared.
The method capital adopts is to seek other quantifiable certainties:
Is there an IPO timetable?
Have top-tier institutions invested?
Is the next funding round already queued up?
How much are listed peers worth?
Thus, risk has actually shifted.
Technology risk hasn't disappeared; it's just temporarily masked by rising valuations.
And when a company's valuation rises from 5 billion to 20 billion yuan, it actually needs to prove more.
A 5-billion-yuan valuation might only require proving technological potential.
A 20-billion-yuan valuation needs to prove customers are willing to pay.
50 billion may require proof that revenue can be sustainably expanded.
Upon entering the hundred-billion-yuan level, the market will ultimately demand answers regarding profits and cash flow.
The earlier the valuation discounts the future, the more performance the company needs to deliver later.
6
The Real Danger Is Not More Decacorns,
But Mistaking "Having a Buyer" for Value
Therefore, the re-emergence of numerous decacorns in the primary market in 2026 does not necessarily signal a bubble, nor can it be simply understood as capital turning reckless again.
AI, robotics, nuclear fusion, and quantum technology do represent important future technological directions, and some companies among them have already seen real orders, revenue, and technological breakthroughs.
The real distinction that needs to be made is:
Is capital buying long-term enterprise value, or a trading opportunity created by a temporarily open exit window?
If a company, after each financing round, can convert the capital into better technology, greater revenue, lower costs, and stronger commercialization capabilities, then valuation increases have an industrial foundation.
But if the main rationale for valuation increases is merely:
Someone invested in the last round;
The next round will be more expensive;
The next step is preparing for an IPO.
Then the so-called "certainty" is essentially just that the transaction chain hasn't broken yet.
Similar stages have occurred in past waves like autonomous driving, computer vision, and the previous semiconductor investment surge.
When the market was hottest, institutions scrambled for allocations; once the financing environment changed, investors ultimately returned to asking the same question:
What exactly has this money bought?
This might be the most important perspective for understanding the "decacorns everywhere" phenomenon in 2026.
The primary market hasn't suddenly become more risk-loving.
Quite the opposite.
As technology and business models become harder to predict, it is now more frantically searching for certainty.
It's just that the "certainty" being chased now has shifted from the enterprise certainly making money in the future to:
A higher probability of being able to list, and a higher probability of someone being willing to pay a higher price in the next round.
These two types of certainty can help investment institutions shorten waiting times, but they cannot substitute for the enterprise truly creating value.
Therefore, regarding the addition of 19 decacorns in embodied intelligence within half a year, what's truly worth attention isn't that China suddenly gained so many 10-billion-yuan companies.
It's what happens in the next stage, when capital once again demands validation through revenue, profit, and cash flow:
How many of today's 10-billion, 20-billion, or even 100-billion-yuan valuations can truly endure.
This is the final exam that this round of "certainty trading" in the primary market must eventually face.
References: First Finance, Shanghai Securities News, Securities Times, Caijing Magazine, Public corporate financing information
Risk Disclosure: This article provides venture capital industry analysis based on public financing and market materials and does not constitute investment advice. Valuations of unlisted companies mostly come from private financing transactions with limited liquidity and are not equivalent to realizable enterprise value; some financing and listing plans are market-disclosed information, and final situations should be subject to official announcements by the companies and regulatory authorities.
This article is from the WeChat public account "BT Finance" (ID: btcjv1), author: Jiang Xu
