Source: Talkie official website

Recently, a statistical chart on the application of AI came out, and the outside world was really surprised.

In Sensor Tower's statistics on the number of downloads in the United States AI app market in the first half of the year, at least three of the top ten are Chinese companies. Among them, the AI education application Question.ai and the AI chat application Poly.ai both come from the "education factory" homework gang, and the other product, the AI chatbot Talkie, comes from the AI large model unicorn MiniMax.

Source: Sensor Tower

Since the AI model ushered in a qualitative change, AI applications have slowly exploded.

In addition, Robin Li, Zhou Hongyi and other Internet bigwigs continue to emphasize that "AI applications are more important than volume models", and domestic mainstream AI players have also begun to shift from volume models and volume technology to "large model price wars", and even developed to crazy "volume" AI applications.

It seems that AI applications have become a necessary weapon for large manufacturers in the AI 2.0 era.

Of course, strategically rolling is one side, and the volume market is the other side of each company's efforts. Players have set their sights on overseas markets.

In the process of competing overseas, the players saw this reality early on: in the wave of AI applications, it is not enough to "roll" in the domestic market, and they must go overseas.

In 2024, AI and going overseas have already become one of the few tracks in the market that still retain increments, and with the blessing of the two major industry dividends of "AI + going overseas", the track has begun to become crowded with the influx of Internet giants to start-up companies and even individual entrepreneurs.

The wave of AI going overseas is becoming more and more turbulent, and when domestic AI applications are involved in "internal injuries", what kind of blue ocean is the global market? How can AI applications go overseas to successfully pan for gold?

01 The three major forces have their own magical powers, and the era of AI exploration has arrived

In the first half of 2024, AI products going overseas have been booming.

The data is confirming the popularity of AI application entrepreneurship and AI products going overseas. According to the statistics of extraordinary production and research, there are about 1,500 AI companies active in the world, 103 Chinese companies going overseas, and the vast majority of AI product startups are concentrated in the application layer, accounting for as much as 76%, as of May this year, there are 1,749 AI products with more than 10,000 monthly visits worldwide, of which 71 are Chinese AI products going overseas.

At present, among the participants in the wave of AI going overseas, there are not only leaders who have made achievements in the mobile Internet era, but also star unicorns who have grown up in the AI 2.0 era, and many start-ups and independent developers have surfaced.

Most of the AI products launched by the "three major forces" are concentrated in areas where generative AI has significantly improved, such as AI chat companionship, AI image generation, AI video generation, AI search, and AI productivity tools.

ByteDance, known as the "App Factory", continues a similar line of thinking when AI goes overseas, as if it has become an "AI application factory", and has also released or upgraded 7 products and platforms overseas in the AI 2.0 era. Among them, Gauth ranked among the top 10 in the App Store and Google Play's free app education list in May after accessing the AI model.

Baidu, which has always wanted to rely on AI to "rejuvenate" in China, also launched three AI applications overseas last year, namely SynClub, WiseAI, and Meira.

Born out of Baidu's homework gang, it is a "big education factory", and the two AI products that have been concerned about going overseas have made some breakthroughs. Among them, Question.AI dominated the list of United States AI education applications in the first half of this year, and the monthly activity of Poly.AI was close to 6 million, and it was ridiculed that "major education companies are not doing their jobs".

As a veteran of going overseas, Wondershare Technology has already released a number of AI products overseas, including Filmora, DemoCreator, and Wondershare Pixcut, which were shortlisted for the top 25 AI traffic in December last year.

As for the large-scale star unicorns, Zero One Everything, MiniMax, and the Dark Side of the Moon have all launched products for overseas.

Among them, MiniMax launched the AI role-playing chat app Talkie in United States as early as June last year. According to Sensor Tower, by July this year, its global monthly active data had reached 11 million.

Zero One is the one that has released the most application products overseas among the domestic large-scale model unicorns. Since last year, four products have been launched overseas, namely PopAi, Monaland, Shado and Bingo.AI.

The dark side of the moon has been revealed to have launched two apps overseas, Ohai and Noisee, although the rumors have been officially refuted, the outside world can still see the connection between these two products and the dark side of the moon.

In addition to the Internet giants and large-model unicorns, there are also many start-up teams and individual entrepreneurs in this wave of AI products going overseas.

For example, the AI education product Answer.AI was launched by the Chinese team, and its team members revealed that the product will be launched in July 2023, and the number of registrations has exceeded one million within 5 months.

There is also FlowGPT, an AI prompt word community, and the founder Dang Jiacheng is a post-00 generation who spent 3 days writing a product demo on his own. Within a week of its launch, the product had tens of thousands of registered users, and as of May, it had surpassed 4 million monthly active users.

In the past year or so, all parties have been staking their ground in the global AI market, and the potential and influence of China's AI are accelerating around the world.

02 AI applications go to sea, nine deaths and one life

AI large models have been popular for nearly two years, and the AI entrepreneurship boom has been surging.

In the past year or so, there has been no shortage of stories of AI "making money". The most typical is HeyGen, an AI video star company, which reportedly grew from $1 million in revenue to $35 million in a year.

The overseas productivity products of 010,000 things are not far behind. Founder Kai-Fu Lee said that the total number of overseas productivity application users is close to 10 million in the 9 months since the product was launched, and the revenue is expected to exceed 100 million yuan this year. According to statistics, the total revenue of MiniMax's Talkie in the first half of the year exceeded 840,000 US dollars (about 6.03 million yuan), with an average monthly income of up to one million yuan.

In addition, on the list of AI products, CrushOn, which was launched in September last year, ranked first for many consecutive months in the first half of this year, and according to statistics, the monthly turnover of the product can reach $60,000. Fotor, which ranks in the TOP3, has increased its followers by 150 million in the past year, and it has achieved profitability early.

It is precisely because of these "wealth-making" stories that are difficult to imagine in China that countless AI entrepreneurs are eager to go to overseas markets.

In the eyes of these AI entrepreneurs who go overseas, going overseas means a broader source of traffic. An AI entrepreneur shared on social platforms that he is very anxious about making AI products in China every day, because opponents not only imitate functions, but also roll up prices and dig corners, but facing the world, the source of traffic is very wide, and you can grab a wave of users by getting a region.

Going overseas means that you can access large-scale model technology with better performance. Byte's overseas version of Coze supports GPT-4o, which has circled a wave of users with a smarter and more powerful chat experience.

Going overseas also means mature user payment habits. Overseas markets are more receptive to the paid membership model, and AI applications can also be quickly monetized by paying users. An overseas report surveyed the pricing models of 40 AI products, and found that 70% of these companies use subscription models.

Going overseas means a more open financing environment. According to data from Viva Industry and Research, AI has attracted more than 20% of venture capital in North America in the past year, and the cumulative financing of the global AI industry has reached 314.9 billion yuan, and the amount of overseas financing is significantly higher than the domestic level.

Although there are many dark horses quietly appearing, more often than not, many AI products going overseas will soon return to dullness, and they will be stillborn before they are even launched.

Taking Baidu as an example, the three products that have gone overseas are not too good at present. Among the 50 products with high visits summarized in the list of AI products in July, Baidu's products are not available.

At the same time, it is worth noting that according to the statistics of this list, among the 50 products, only the overseas version of Kelin, which has just been launched, has seen a significant increase in visits, most of the products have performed steadily, and even the number of visits to many products is declining.

AI Product List

Among them, CrushOn, PixAI.Art, HIX AI, Coze, CiciAI and other products have seen a year-on-year decline of more than 30% in terms of visits.

In general, the success rate of AI products going overseas and achieving wealth creation is not high, and there are very few that can become popular and operate stably.

AI products that can achieve "wealth" are often inseparable from team members' insight into user needs, the accumulation of AI technology, and the control of innovation, but the products that fail are different behind them.

An AI practitioner shared on social platforms that he tried to do three AI projects in 2023, but none of them came out, some because the homogenization was too serious, and some because they chose the wrong direction at the beginning, and all of them ended in failure. Another AI practitioner said that many AI products end in the cold start stage. These products have not done user research, and are more enthusiastic about imagining usage scenarios out of thin air, lacking growth strategies and operation strategies, and it is difficult to really run through.

In the broader global market, it may be easier for products to scale up and increase revenue, but entrepreneurs also face more brutal competition.

03 The global market seems to be vast, but it is not "gold everywhere"

In the long run, it is an inevitable trend for AI to go overseas, and even many companies are based on globalization at the beginning of their start-ups.

As Zhu Xiaohu said, AIGC is a long slope and thick snow in the next 10 years, AIGC will redo all software, consumer electronics and consumer applications, and the application layer will create the most value. Kai-Fu Lee has always believed that 2024 is the year of application explosion, and in the AI 2.0 era, all applications are worth being rewritten.

In the global market, AI has long developed into a "gold mine". According to research reports, the size of the global artificial intelligence market will be about $454.12 billion in 2022, increase to $538.13 billion in 2023, and is expected to reach $638.23 billion this year.

However, the development of AI is still in its early stages. Cautious people even believe that AI is far from entering the "mobile Internet era", and the qualitative change in the technical capabilities of AI large models, just like the Internet era, is still at the level of improving infrastructure, and is still incubating the environment for super applications, and there is still a distance from the emergence of killer applications.

At this stage, there are many people rushing into the global AI market, but AI products are still in the reckless period when they go overseas, and it is not easy to tap this gold mine.

First of all, AI products going overseas must make the target users truly feel the use value, so as to quickly increase the volume and retain users.

Insufficient understanding of the ecology of overseas users, inaccurate grasp of the needs of overseas users, and inability to achieve effective product innovation are important reasons why it is difficult for many AI products to develop overseas. When the homogenization of AI products becomes more and more serious, only AI products that truly make users feel valuable to use and truly run through the closed-loop can allow users to retain and pay.

Second, commercialization is a challenge that all AI products have to face.

OpenAI, which once set off the source of the wave of AI change, is not having a good time now. Not only has growth stalled, but some analysts say that the annual loss is as high as $5 billion.

In addition to OpenAI, retaining users, opening up the business closed loop, and achieving profitability are still problems for many AI products going overseas.

Even some people who have already made money are not in a very good situation. Because many independent developers have earned their first pot of gold in the form of AI shell applications. In the past, the explosion of many AI products was due to the time window after the qualitative change in the technical capabilities of AI large models, and to circle a wave of users through micro-innovation, which was also described as "small volume, light assets, and making quick money".

However, to make AI products go overseas, it is difficult for independent developers to fight for the "regular army" of forming a team, getting financing, and going public. When the "regular army" becomes stronger and stronger, the living space of such products will be eroded, and eventually it will become smaller and smaller.

In addition, compliance is a difficult problem that all enterprises have to face.

According to the Beluga report, recently, some payment service providers for AI overseas products have stopped cooperating because they are worried about risks, resulting in difficulties in product development, and AI character chat and AI pictures are the two key categories affected.

On August 1, the EU Artificial Intelligence Law came into effect, putting forward higher requirements for Chinese AI enterprises going overseas. The world's first comprehensive regulation of artificial intelligence, which requires AI-generated audio, video, images, and chatbots, is punishable by fines of 35 million euros or 7% of annual global sales.

Volume, commercialization, and compliance are like the three mountains of AI companies going overseas, and only by crossing these three mountains can it be possible to truly "pan for gold".

There are many chasms that need to be crossed when China's AI goes overseas and makes the next "TikTok".

This article is from Xinzhi self-media and does not represent the views and positions of Business Xinzhi.If there is any suspicion of infringement, please contact the administrator of the Business News Platform.Contact: system@shangyexinzhi.com