Does the so-called "first publicly traded AI video stock" really matter?

Who will become the first public AI video company seems to be a new suspense in this track. 

This issue appears to be related to industry rankings, but in essence, it first concerns the exit of investors. The first company to go public will get the chance to obtain scarcity valuation first, open public financing channels, and let early shareholders see the possibility of realizing returns. 

Especially when the capital window may change at any time, the successful listing of the first company may also mean that latecomers lose the opportunity to go public under the same conditions. 

Therefore, the first public AI video company is obviously very important to investors. But for an enterprise to make good products, retain customers and build a sustainable business, listing itself may matter, but being the first to go public is not that important. 

Among the comprehensive AI enterprises that have been listed so far, there are no shortage of companies with video generation capabilities. 

However, those who are truly qualified to compete for this title should be enterprises that take AI video as their main product, commercialization direction and capital story. 

Shengshu Technology has completed the shareholding system reform, Aishi Technology and Yanyu Technology have successively spread news of listing in Hong Kong, and Keling has also been given extremely high independent listing expectations after completing independent financing...... Around these companies, the competition for the first public AI video company has become more and more specific. 

What is really worth discussing is not who is most likely to ring the bell first, but why these companies that have only been established for two or three years are all rushing to the stock market.

01、The competition for "the first public company" is for the first-mover advantage, and what investors fear is having no fallback options

Over the past year, the financing scale of AI video companies has risen significantly. 

In July this year, Aishi Technology announced the completion of the extended financing of Series C round, and the total financing amount of the entire Series C round reached 439 million US dollars; in the same month, Shengshu Technology completed a new round of 500 million US dollars in financing. By the end of August, the National Artificial Intelligence Industry Investment Fund agreed to invest 1.4 billion yuan in Beijing Keling, and the total subscription amount of Keling's financing in this round reached about 20.447 billion yuan. 

Younger teams such as Sand.ai, Zhixiang Future, and Flova are also completing financing rapidly. From the underlying model to the Video Agent, and then to the creation platform that aggregates different models, capital has almost bet on every possible route. 

Screenshot of video generated by Sand.ai

The influx of capital shows that AI video is still highly sought after, and it also pushes a more realistic issue to the forefront: how will all this capital exit in the end?

For companies that have just completed the seed round, IPO can still be a long-term plan. 

For enterprises that have completed consecutive rounds of financing and their valuations have entered the unicorn stage, listing can no longer stay in the imagination. After completing each large round of financing, the company obtains more development funds, and also accumulates higher shareholder return expectations. 

The primary market can continue to push up valuations through the next round of financing, but this process cannot last indefinitely.

When the financing scale has reached hundreds of millions of dollars, who will take over the next round, who is willing to accept a higher valuation, and when early investors will exit will all become more and more difficult to answer. 

Keling's financing arrangement provides a very telling detail. 

According to the announcement of Kuaishou, some new investors have obtained redemption rights at the same time. This does not mean that investors will definitely demand to exit, but it shows that even for top projects like Keling, capital will arrange retreat routes in advance when investing huge sums of money. 

Listing is exactly the most easily understood answer. It may not allow shareholders to cash out immediately, but it can establish public trading and subsequent financing channels. At the stage when the industry still lacks similar listed companies, the first enterprise still has the opportunity to obtain scarcity and establish a valuation coordinate according to its own growth logic. 

However, after the first company goes public, it will also become a reference for latecomers.

If it performs well, peers can continue to tell similar growth stories; if it performs poorly, latecomers may face more cautious investors, lower valuations and stricter listing requirements.

The experience of Unitree after listing has shown how fast this change can happen. 

In August this year, Unitree closed up 460% on the first day of listing on the A-share market, and then its stock price fell significantly. Recently, there have been news that the regulatory authority has raised the listing requirements for humanoid robot enterprises through informal window guidance, focusing on sustainable revenue, loss reduction progress and substantive innovation. 

Unitree's stock price trend

The Unitree case took place in the A-share market, while the listing plans of AI video companies are mainly targeted at Hong Kong. Essentially, it is also because the A-share listing itself has a higher threshold. Robotics or embodied intelligence is "hard technology" to some extent, but at present, AI video generation, which has more exit options, obviously can't even be called "soft power". 

Such signals are likely to further intensify the mentality of AI video companies to compete for the first place.

Embodied intelligence is more directly connected with manufacturing upgrading and the real industry, and it still needs to answer whether the revenue can be sustained, whether the loss can be narrowed, and whether the technological innovation is sufficient to support the valuation. 

It is even more difficult for AI video enterprises to continuously obtain capital patience only through grand narratives such as "world model" or "next generation content production method".

Especially for those companies that have completed multiple rounds of financing but cannot survive by their own cash flow, once the IPO expectation fails, the management must explain to investors: who will pay for the next round, how to maintain the existing valuation, and how long shareholders have to wait.

Therefore, "the first public company" is first and foremost a competition about exit order.

What everyone is competing for may not be who is the first to prove success, but who can give investors an explanation before the window narrows. 

02、Ringing the listing bell is not the end, only customer payment can bring a future

Investors' anxiety about the listing order cannot be directly transformed into the product advantages of the enterprise. 

Being the first to go public will not automatically bring better models, more stable generation effects, nor will it make users more willing to renew their subscriptions. 

The capital market can determine how many resources a company has temporarily, but the product market still needs to test what these resources produce. 

Companies that develop their own basic models need to continuously invest in algorithms, data, computing power and talents. They must prove that the improvement of model effects can be transformed into sufficient subscription and call revenue, and retain reasonable profit margins after deducting the high training and reasoning costs. 

This is obviously not easy. Video models are still iterating rapidly, today's leading generation effects are likely to soon become the industry standard.

Source: Artificial Analysis

Large manufacturers can also use cloud services, traffic entrances and product ecosystems to lower prices. If a startup can only keep increasing investment but cannot establish stable customer relationships, the more financing it gets, the higher the growth that needs to be realized in the next stage will be. 

Video Agent and application platforms face another set of problems.

Nadou Pro and Pollo AI can aggregate multiple models, and Medeo can build workflows around conversational creation. But when competitors can also call the same models, what can really retain users can only be higher creation efficiency, more mature workflows, and long-term accumulated customer relationships. 

For people who actually use the product, the judgment criteria are more specific. 

Advertising agencies or short drama producers will care whether the same character can remain consistent across shots, how many times a finished video needs to be generated, how much manual revision is required, and whether the final cost is really lower than traditional production. 

There is still a long way to go between generating a stunning picture and stably completing a whole work. Only by doing these links well can the generation ability be transformed into customer repurchase. Being the first supplier to go public will not make customers accept higher costs and more unstable delivery.

Screenshot of AI-generated video

Ordinary creators will not choose products according to the order of ringing the bell either. A company that goes public later but has a more user-friendly product, more suitable price and more reliable generation results can be switched to by users at any time. 

The conversion cost of current AI tools is not high in the first place, and whether a company is listed is even less likely to constitute a user barrier.

Financing advantages can certainly be transformed into operational advantages. More funds mean that enterprises can buy more computing power, train larger models, and have the conditions to bear longer trial and error time. 

But this transformation requires specific organizational and operational capabilities. Funds may only help an inefficient company maintain its original outdated business model for a longer time. 

Therefore, to judge whether an AI video company has truly established its business, we ultimately need to see how long paying users can stay, whether enterprise customers renew their contracts, how much reasoning and customer acquisition cost corresponds to per unit of revenue, and whether operating cash flow has improved after R&D investment continues to increase.

For enterprises with consecutive rounds of financing, what is more important is whether after each round of capital investment, it has accumulated more stable customers, improved operational efficiency, and moved one step closer to surviving on its own revenue. 

Source: Internet

The real industry value of the first listed AI video company may also lie here: it will put the revenue structure, cost pressure and customer quality of such enterprises in front of the public market more completely, giving the outside world the first chance to test what kind of business AI video is on earth.

Perhaps many of these companies are more worried that being the first will become the "only". After all, the voices of bearishness on the AI bubble are getting louder and louder now, and as mentioned at the beginning, the AI video track is obviously not a product with real transformative power, and the peripheral fields like this will naturally be the first to be affected when the bubble is squeezed. 

Enterprises can certainly strive to be the first to go public, because they need to give an explanation to investors or their own entrepreneurial ambitions. 

But after ringing the bell, it still has to answer a more difficult question: can this company make products that users are willing to pay for in the long run or that are valuable? 

This article is from the WeChat official account "Yiyu Observation" (ID: yiyuguancha), written by Yishu Team, and authorized for release by 36Kr.