Large models will now need to seek funding from the secondary market.

Why Are AI Large Model Companies Trapped in a Cycle of Capital Consumption?

**Produced by｜Huxiu Technology Group
**

Author｜Song Sihang

Editor｜Miao Zhengqing

Header Image｜Visual China

Just this week, market news indicated that Anthropic and Dark Side of the Moon are both expected to enter the capital market in 2026.

The former is valued at about $380 billion and has advanced to a Series G funding round; the latter is valued at approximately $180 billion, making it one of the highest-valued model companies before an IPO in China. The foundational models developed by both companies have consistently ranked at the top of various technical community lists.

At the same time, OpenAI’s listing plans have frequently been mentioned, and Jumping Stars is also rumored to be in the preparation stage.

This means that in China and the United States, the two countries with the strongest large model capabilities, leading foundational model companies are almost all heading towards the capital market around 2026.

Large model companies are collectively moving towards the same “finish line.”

There are two signals behind this trend.

First, this round of large model competition has shifted from a focus on capabilities to validating commercial value.

In the past two years, the industry was more concerned about who was closer to SOTA and who could take the lead in the next model release. But now, the focus has changed; with the popularity of Agent applications like OpenClaw, people are beginning to pay attention to whether the capabilities of large models can be sustained through continuous investment and whether they can convert into verifiable commercial value.

The second signal is that once model companies reach a valuation of hundreds of billions of dollars, the patience of the capital market has also reached its limit.

As of now, if these large model companies claiming to go public in 2026 successfully list, then by the end of 2026, almost all leading large model companies in China and the United States will have converged in the secondary market.

When leading companies start intensively moving towards IPOs, it indicates one thing: this track can no longer rely on the patience of the primary market to maintain operations. Large model companies still need money. However, capital is becoming less willing to wait indefinitely.

Over the past two years, the large model track has siphoned off a significant amount of funds. Whether in China or the United States, leading companies have raised tens of billions or even hundreds of billions of dollars. This money was seen as a “bet on the future” at the time, but now it has begun to turn into accounts that need explanations.

If we calculate from the end of 2022, when ChatGPT appeared, it has already been three years. In this three-year period, the valuations of large model companies have skyrocketed to over a hundred billion dollars. This pace is exceptionally fast compared to any previous era.

So why do large model companies need so much money? And why can the primary market no longer afford to provide these funds?

Remember when large model companies first emerged? No one regarded them as “infrastructure” back then; the infrastructure at that time was computing power. But now, with AI applications flourishing, large models have also become “infrastructure.”

If we value them as “infrastructure,” they are no longer light asset businesses.

The essence of infrastructure is continuous investment. Whether it’s model training, inference costs, or the ongoing iterative upgrades to maintain performance leadership, all point to one thing: these companies cannot establish barriers through one-time investments but must continuously consume resources.

In other words, large model companies are entering a long-term capital consumption structure.

This also directly brings about a more realistic problem: their revenue currently cannot support this structure. Whether it’s API calls, enterprise clients, or subscription models, while all are growing, the growth itself is not stable, and prices are continuously declining.

The stronger the model, the more calls it receives, and the higher the costs; yet price competition is continually compressing profit margins. This is a typical case of “the larger the scale, the greater the pressure.”

Therefore, the demand for funds from large model companies is not temporary but continuous. The problem is that this demand has begun to exceed what the primary market can bear.

The primary market can bear the risks for a company, but the premise is that this risk has a clear time boundary. However, the biggest problem for large model companies is that—time boundaries are not clear.

They are neither like internet platforms that can quickly generate network effects through user scale, nor like traditional infrastructure that can gradually recoup costs through stable cash flow.

They find themselves in a more awkward position: needing to invest at an “infrastructure” scale but lacking the “infrastructure” return capability.

When a single company requires continuous investments of billions or even hundreds of billions of dollars, while the return cycle remains uncertain, the logic of the primary market begins to falter.

From this perspective, when long-term capital is no longer willing to unconditionally extend life, financing is no longer an action that can be repeated indefinitely.

In this situation, large model companies have only one choice left: to go to a place where they can continue to talk about the “future,” which is the secondary market.

It has been proven that the market sentiment towards AI is already sufficiently optimistic.

Take MiniMax as an example; it opened high on its first day in the capital market, surged during the session, and experienced several noticeable increases in the following trading days, with its market value consistently rising compared to its issuance stage. Similarly, “the world’s first large model stock,” Zhipu AI, also showed a comparable performance, experiencing brief fluctuations in the early stages of listing, followed by a rapid rebound in stock price, with some trading days recording single-day increases of over 8% or even 10%.

If we pull the timeline further back to compare with the previous generation of AI companies, such trends are quite rare.

Whether it was Mobvoi or SenseTime, both experienced a prolonged valuation digestion period after going public. Stock price performance was more cautious, with market doubts regarding their commercialization capabilities long suppressing valuation space.

So, what lies ahead for OpenAI, Anthropic, Dark Side of the Moon, and Jumping Stars, these large model companies about to go public, is to wait for the capital market to provide the final answer.