AI investment makes entrepreneurs sad

As AI tools become more popular, how to attract and retain users has become a major challenge for AI startups. However, the high cost of acquiring customers and fierce market competition have put a lot of pressure on many AI entrepreneurs. This article will explore the dilemma of AI startups in the battle for traffic, and how they can find new ways out in the ever-changing market environment.

Sam Altman, founder of OpenAI, believes that the AI ​​era will create a new type of startup, where one person can build a billion-dollar unicorn company.

Now, this era has not really arrived yet, and AI entrepreneurs are blocked from the door by the high investment costs.

In the words of AI entrepreneur Zhao Tian, ​​"In 2023, everyone is talking about technology, products, and scenarios. In 2024, it will be about two things: growth and making money."

Zhao Tian used to work for a large domestic company. He started his own business last year. His products are related to AI image generation. The early stage of his business went smoothly and he successfully completed the A round of financing. Since October last year, he has done some exposure-type promotions on Bilibili, Xiaohongshu, and Douyin, with considerable conversion effects. The CPM (Cost Per Mille) is about 5 yuan.

"I dare not say that our products are good, but in 2023, users are full of curiosity about AI tools and are willing to try them. As long as they are exposed to the right people, there is a high probability that conversion will be completed," said Zhao Tian.

However, with the high-profile entry of the Dark Side of the Moon, the threshold of this game, which was originally a game of small bets for big gains, suddenly became unattainable. The products of unicorn companies and Internet giants have absorbed most of the traffic and attention, and have pushed the price of investment to an unprecedentedly high level. Ordinary entrepreneurs can either grit their teeth and keep up, or leave the market early.

They invest in advertising because of anxiety, and investing in advertising in turn fuels anxiety. No one knows whether every plan and every confirmation in the advertising bidding system will make them live longer or get closer to death.

01｜“I haven’t seen such a generous sponsor since Pinduoduo”

Kimi is the most popular AI product after the MiaoYa camera. Unlike most short-lived "Internet celebrity products", Kimi has always maintained a high popularity after its release.

In March this year, the monthly visits of Kimi under Dark Side of the Moon exceeded those of Baidu Wenxin Yiyan, becoming the first in the monthly active and visit volume of ChatBot. Half a year later, this leading position has not changed. According to the August data of AI product list aicpb.com, Kimi's monthly visits reached 23.36 million, still ahead of Wenxin Yiyan's 19.16 million. Minimax, a star startup company of large models like Dark Side of the Moon, has only 513,900 visits to its Conch AI.

Naturally, Kimi's ability to attract traffic and generate user fission has also become the subject of research by various AI companies.

It is almost universally recognized in the industry that Dark Side of the Moon is the most aggressive large-scale startup in advertising. According to data from AppGrowning, Kimi spent 15 million yuan on advertising in March this year, and nearly 50 million yuan in July and August, with a total investment of 140 million yuan in six months.

Screenshot of the Dark Side of the Moon official website

Station B has become the "home" of the Dark Side of the Moon. A UP host told "The Mountain" that he can only accept two business orders per month, and for two consecutive months, all the business orders were given to the Dark Side of the Moon. As far as he knows, not only the technology area, but also the workplace, knowledge and other fields have waited for the gold coins scattered by the Dark Side of the Moon. "I haven't seen such a generous sponsor since Pinduoduo."

In addition to content cooperation, Kimi also frequently places information flow advertisements directly. Some media quoted multiple practitioners as saying that Kimi's CPA (Cost Per Action, user conversion cost per person) quotation on Bilibili is as high as 30 yuan, while at the beginning of this year, the price was about 10 yuan.

The Dark Side of the Moon has been flooded with ads, and other companies are quickly following suit. On Bilibili, Doubao bought a competing entry, and searching for Kimi will result in Doubao ads. On Xiaohongshu, many people are teaching you how to use Tencent Yuanbao to plan travel itineraries. Podcasts, which were previously seen as niche channels, are also increasingly being heard of by AI tools.

The advertising war has continued from online to offline. There are already Zhipu Qingyan ads in Beijing subway stations; Minimax has started to put ads in elevators in Lujiazui, Shanghai, where white-collar workers frequent; Doubao, like Feishu, has also started to tell people how to "use AI to drive business growth" at airports.

Companies are also actively embracing these AI sponsors. At the second quarter earnings conference in 2024, Chen Rui, CEO of Bilibili, will take the initiative to mention that Bilibili has high-quality AI content; Xiaohongshu has also increased the operation of AI content this year, and some ordinary operation activities will receive direct support from the company's product and technology executives; Focus Media has also begun to increase sales efforts for AI companies, and finally persuaded some unicorn companies to place elevator advertisements.

With hundreds of investment, traffic has also risen. According to Similarweb statistics, the total visits to the products of the five AI unicorns (Zhipu AI, MiniMax, Baichuan Intelligence, Zero One Everything, and Dark Side of the Moon) have skyrocketed by 963% in 6 months.

02｜Knowing the rules is one thing, copying the next Kimi is another

Zhou Xian is in charge of product operations at an AI startup company related to the second dimension. In the past, her work mainly involved product planning and design, expanding partners and channels, etc. This year, her work is almost limited to one thing: investment and growth.

“In almost every regular meeting, we first talk about Kimi’s recent investment, and then talk about our own investment growth.” Zhou Xian said, “Growth has become the most important thing for the company.”

Zhou Xian said that the growth pattern is not difficult to summarize. To put it bluntly, it is to invest a large amount of money in the simplest and crudest way to quickly occupy the minds of users and then achieve downloads and conversions.

However, knowing the pattern is one thing, and copying the next Kimi is another.

Kimi uses a CPA (cost per action) traffic model, which means that when a user completes a corresponding action, usually registration or downloading, Dark Side of the Moon needs to pay a fee. This is a traffic strategy that can directly drive growth, and it is simple and crude.

Traffic investment bidding is similar to auctioning. If others bid too high, then the people who bid low will get less traffic. Well-funded Internet giants and large-model unicorns raise the cost of traffic investment, making it difficult for ordinary entrepreneurial companies to follow suit.

Zhao Tian said that entrepreneurs like him have no way of affording a CPA cost of 30 yuan. He did some calculations and found that early-stage AI companies usually get only 3 to 5 million yuan. "Let's not talk about anything else. If we take out all the money and invest in advertising, we will have at most 200,000 users, and there is no guarantee of retention."

As for Douyin, it has also become an "AI forbidden area." Zhou Xian said that except for ByteDance's own Doubao or AI products with investment and cooperation, traffic will be restricted, either because the content is "illegal" or the traffic consumption is slow - which means that your content is not watched by many people.

Doubao’s advertisement on TikTok

As the cost of traffic increases and the novelty of AI decreases, the cost of acquiring customers for AI companies is also increasing. Some practitioners said that Kimi's total cost of acquiring customers is now about 15 to 20 yuan. According to previous estimates by Sina Technology, Kimi's daily customer acquisition cost is 200,000 yuan. This means that not only will Dark Side of the Moon have to pay increasingly high growth costs, but other AI companies will also have to get involved in this advertising bidding and be forced to increase their traffic costs, which makes many entrepreneurs discouraged.

03｜The unsustainable investment war

The war has entered a white-hot stage, and the unicorns and Internet giants that first launched the charge have also begun to worry about the increasingly high cost of acquiring customers.

Kimi is the only product of Dark Side of the Moon. Its founder Yang Zhilin has also said on many occasions that they currently do not have a next product and want to gradually turn Kimi into a "super application."

In the era of mobile Internet, creating a "super app" is the ultimate dream of many entrepreneurs. Eventually, applications such as WeChat, Douyin, and Meituan were born, which are all-encompassing in their respective fields and have the ability to expand to other fields. Companies have also become giants of the new era based on this.

To become a super app, two basic standards are essential. First, the number of users must be large enough and they must be used frequently enough. Second, these users will not easily migrate to other platforms, that is, to build a high enough moat.

On the surface, traffic investment is to solve the problem of the number of users. However, the improvement of user indicators such as DAU cannot fully enhance the competitiveness of the platform.

A product manager of Dark Side of the Moon, “Song’e”, posted on the social platform Jike that high DAU does not necessarily crush other competitors. “AI provides intelligent value, not directly information exchange value. Under most current product designs, there is no multilateral network effect, and the growth of DAU will not bring exponential value improvement to the product.”

However, although the increase in the number of people may not bring economies of scale to Kimis, it can help Kimis obtain data and feedback in a more natural environment, thereby improving the product experience.

Yang Zhilin positioned Kimi as a "super assistant". He said: "I think in the AI ​​era, the super application will most likely be an assistant. I think the demand for intelligence is a very universal demand, but today its capabilities are still at an early stage."

An important step in training a "super assistant" is to have enough data. Yang Zhilin believes that data should not be a fixed data set, but a "variable". "How to use data, or get user feedback, will become more and more important."

With this idea, acquiring enough users is also a way to build a data moat. However, for a commercial company, any strategy must measure ROI. According to statistics from aicpb, in the past three months, the average number of days that users used Kimi was four days, and the average usage time was no more than two minutes. It is difficult to say how much data feedback such interaction can bring.

Companies have also begun to reduce the cost of traffic. 36Kr reported that Kimi's high marketing costs have made some shareholders feel unsustainable. An employee of the Dark Side of the Moon said that in the fourth quarter of this year, the company will try to reduce the cost of traffic and increase the investment in new channels such as Video Account, Weibo, and Zhihu.

On these platforms, Kimi's related materials are more "down-to-earth", with titles often like "Use Kimi to get off work early", "Use Kimi to finish Excel in ten seconds", "Use Kimi to become a self-media blogger with a monthly income of 100,000", etc. Some of these materials come from material service providers, and a large number are produced by Kimi's market growth department itself, and some employees will also appear in person.

It is hard to say how many truly valuable users these materials can attract. When users are attracted by these exaggerated titles, they find that AI cannot solve their problems once and for all after actually using them. Many bosses of small and medium-sized enterprises who pay attention to cutting-edge technology have tried it and finally found that manual labor is more convenient and cheaper than AI.

Some investors predict that the money-burning war may end within half a year. By then, unicorn companies may not have much money to continue investing, and large companies will also need to evaluate ROI. Only then will everyone gradually become more rational, paying more attention to commercialization and paying more attention to users who are willing to pay.

04｜Finding other solutions

Zhao Tian said that he was now in a dilemma. If he did not continue to invest, it would be difficult to achieve sustained growth and obtain the next round of financing; but if he continued to invest, it might only make him run out of funds faster.

Zhu Xiaohu, founder of Jinshajiang Venture Capital, has always been skeptical about AI's C-end products and is even more firmly opposed to "burning money to buy users." In an interview, he said: "It is very difficult to burn money to buy users today. In the past, we could raise three rounds of financing a year, but now it may not be easy. The financing environment is completely different, so we must not burn money."

Starting from 2024, all investment institutions have tightened their investment in AI startups. An investor told "The Mountain" that products with clearer paid conversions, especially B-end products, may still get some money, and C-end products have almost no chance of getting A round or above financing at this stage. As for new large-scale model companies, "the qualifying round is over, and the elimination round starts early." Zhu Xiaohu is even more outspoken, believing that the only way out for the current large-scale model unicorns is to sell them to large companies.

Entrepreneurs like Zhao Tian are not unaware of the cruelty of the situation, nor are they unwilling to look for entrepreneurial opportunities with better paid prospects. However, when they invest in AI startups, both entrepreneurs and investors are more patient, willing to wait for the market to mature, willing to make a small and beautiful product, and willing to tolerate temporary losses.

In just one year, when Zhao Tian went to discuss financing again, everyone told him that he either had to have a sufficient user base to verify market feasibility or be able to immediately prove his commercial prospects. Proving commercial prospects meant almost starting a project from scratch. The only way left for Zhao Tian was to invest in traffic and acquire customers.

More large companies have also realized that it is difficult to retain users brought in by investment traffic when it is difficult to widen the gap between technology and products. Doubao has the advantage of Douyin and can "occupy the minds of users" at a lower cost. Now their work focus has shifted to sales rather than marketing, and they have begun to customize technical solutions for B-side customers.

Ali Tongyi also only cooperates with more precise channels. An Alibaba insider told "The Mountain" that they think AI should be more like a cloud service, so it should be targeted at "producers" rather than just an entertainment application.

Zhao Tian said that in the current environment, it is too difficult to make a "small and beautiful" AI application. He plans to focus on overseas marketing, where the cost-effectiveness of marketing is higher. The cost of acquiring customers on Google and Twitter is only about half of that in China, and people are more accepting of a small product, rather than having a large number of users to obtain financing.

The fear of missing out on AI hangs over everyone. Many people don’t know why they do this, but if they don’t do it, they may not be able to stay at the table. Perhaps both entrepreneurs and investors need more patience. After all, AI is still a child. The technology, products, talents and markets needed for growth all take time and cannot be forced to mature.

"In fact, AI sells services, so as long as you serve a certain number of people well, you can survive. The cost of finding such people in China is too high, but it is much better abroad," Zhao Tian said. "There is no one there to urge you to grow a hundred times in a short period of time."

(Zhao Tian and Zhou Xian are both pseudonyms.)

Author｜Qi Ming Editor｜Xue Xingxing This article is written by the operator author [山上Hillvue], WeChat public account: [山上], original/authorized to be published in the operator, and any reproduction without permission is prohibited.

The title image is from Unsplash, based on the CC0 protocol.