AI companies flock to go public, some soaring and some falling endlessly, but Sibichi is particularly unlucky.
As one of the earliest AI companies in China, Sibichi has been around for nearly 20 years. In 2022, it launched a challenge to the Science and Technology Innovation Board, but was unsuccessful. At the beginning of this year, it planned to launch a second IPO, and the current review status is “inquired”. At this critical moment, three chip dealers reported under their real names that Sibichi and its holding subsidiaries engaged in non compliant and unethical operations.
According to reports from dealers, Sibichi deliberately exaggerated the market prospects to induce dealers to hoard large amounts of goods in order to sprint for revenue scale, and confirmed sales revenue in advance without completing the complete product delivery loop.
They themselves have immature technology, seriously underestimated the difficulty of chip landing and delivery, and even fabricated cooperation resources. The promised mobile and broadcasting operators, as well as top home appliance customers such as Changhong and Konka, ultimately did not land, “said the person in charge of one of the distributors.
For Sibichi, whose business fundamentals were already not very good, the reports from distributors can add insult to injury. If the content of the reports is confirmed, this listing plan may be destroyed by Sibichi’s “big cake” drawn for distributors.
Technology is not good, can only draw cakes?
The growth of Sibichi was originally an inspirational story born in the “cold winter” of artificial intelligence and the first to break the gap in domestic voice technology. However, due to its early entry into the market, it once became the darling of capital. But it woke up early in the morning, rushed to the evening gathering, and as more and more latecomers took advantage of the power of capital to take off, Sibi Chi seemed to lack some momentum.
If the dealer’s report is true, then we seem to be able to uncover the clues behind this weakness.
According to reports, the companies that initiated real name reporting this time are Shenzhen Weihefeng Semiconductor Co., Ltd., Lianxin Semiconductor (Shenzhen) Co., Ltd., and Shenzhen Donghengsheng Technology Co., Ltd., all of which are long-term cooperative distributors of Sibichi. The person in charge of Weihefeng said, “At first, they drew a big blueprint for us, claiming that the technology was mature and the chips did not have to worry about sales. We received 5 million yuan worth of goods in our first order, but after the goods arrived, quality control problems emerged one after another
Multiple distributors have also confirmed that after receiving full payment and delivering hardware chips, Sibichi unilaterally withdrew its technical team and stopped providing core keys, resulting in a large number of sold chips becoming unsold electronic waste.
For a long time, Sibichi has emphasized that the company has the ability to independently develop chips, and both the company’s dedicated chips and general-purpose adaptive chips are independently developed and designed. But now dealers have raised doubts, claiming that the core source of the Sibychi voice chip circulating in the market comes from two external manufacturers in Zhuhai and Shanghai, and some products are even shipped directly without polishing or removing the original factory silk screen.
According to the distributor, confirming channel sales revenue without completing the complete delivery loop has already crossed the red line, involving the exaggeration of book revenue scale, which may directly lead to the failure of listing. But more importantly, whether it is quality control, inventory backlog, or dealer doubts about whether its chips are self-developed, it seems to directly point out that this AI “veteran” lacks technical capabilities in speech recognition and interaction, which is undoubtedly tearing apart the technical label on Sibichi.
What if the technology is not good but revenue growth is needed? Sibichi not only paints a cake for dealers, but also paints a cake for the capital market.
In the last prospectus, Sibichi gave a very optimistic performance forecast, with revenue of 452 million yuan, 725 million yuan, 1.15 billion yuan, 1.69 billion yuan, and 2.451 billion yuan from 2022 to 2026, respectively, with a compound annual growth rate of over 51.46% from 2021 to 2026. In fact, from 2023 to 2025, Sibichi’s revenue will be 539 million yuan, 601 million yuan, and 688 million yuan respectively, with a three-year compound growth rate of only about 12.9%, which is far from the previous performance commitment.
It is precisely because of this unfounded optimism that the company was deemed to “not meet the issuance conditions, listing conditions, or information disclosure requirements” and the listing process was terminated.
On Zhihu, someone once asked a question: What is the experience of working at Sibichi? One of the replies stated, ‘Leaders talk about their industry dreams all day long… even those who openly refuse can leave Sibichi.’ Another sarcastically stated, ‘We are all young people who work hard for our ideals (the boss’s), talking about money.’.
Whether internally or externally, perhaps Sibichi really needs to reflect.
The better the story is told, the more likely it is to overturn
A few months ago, an article in Silicon Valley titled “99% of AI start-ups will die in 2026” was brushed off the screen. Srinivas Rao, the author, said frankly: “The current AI boom is just a copy of the Internet foam. ”
At that time, the Internet was in the ascendant, and a wave of Internet entrepreneurship emerged in the world. Later, although it was proved that the Internet had brought unprecedented prosperity, many enterprises were eliminated because of capital, business model, competition and other problems before the technology was truly mature. This scene will naturally repeat itself in the era of AI, especially for some companies that like to tell disruptive stories first and then talk about monetization, which can now glimpse their future.
When it comes to Wenshengtu AI tools, many people may still remember Stability AI. In 2022, this London startup became the center of the AI craze with the explosive popularity of its AI model Stable Diffusion. After two consecutive rounds of funding, its value skyrocketed from $100 million to $1 billion. At a highly anticipated press conference held in San Francisco, founder Mossack confidently told the audience, ‘We haven’t done much, but I think we’ve overturned the world of AI.’. ”
However, he has not yet disrupted the world of AI and has already “disrupted” his own company.
Mostak is a person who loves to brag excessively. He often speaks eloquently about partners and projects before achieving actual cooperation with Stability AI, and shows a firm attitude towards these collaborations. Even claimed to have contacted prime ministers/prime ministers’ offices in multiple countries to discuss developing AI models for each country.
Within the company, Mossack often makes “bizarre claims” and “lofty promises” that are not in line with the actual situation of the company, but fails to fulfill them.
The high opening and low closing of Stability AI exposes the large loopholes in the governance structure of start-up companies that develop too quickly under the “ripening” of capital. If these companies encounter another manager who is full of nonsense, the only thing waiting for them may be elimination. It is reported that Stability AI has faced severe cash flow disruptions and debt crises in 2025, and ultimately had to undergo restructuring and introduce external capital controls.
In China, a new wave of IPOs is shaping a feast for AI companies, with big model nouveau riche represented by Zhipu AI and MiniMax showcasing high growth models and becoming the most dazzling presence in the capital market. And the excitement belongs to the new aristocracy, while the bleakness leads to the “old people”. Faced with the popularity of AI concepts, the once highly anticipated “Four Little Dragons” seem to have not received any dividends.
This is because the AI nouveau riche, built on the foundation of big models, are using technological upgrades to realize the story originally told by the AI “Four Little Dragons”, while the AI “Four Little Dragons” are heading in a completely opposite direction.
Around 2016, the AI venture capital boom emerged amidst multiple hot AI events, and a large amount of funds began to flood into the industry. The “Four AI Tigers” quickly became stars in the industry. Although they entered different tracks, the early four AI dragons all used “universal AI platforms” as their selling point, claiming to be able to achieve various cross scenario applications through deep learning. It has to be said that the entry of this AI story is far more imaginative and attractive than concepts in single fields such as smart homes and smart transportation.
However, when it comes to realizing the story of a “universal AI platform,” the style of these AI 1.0 stage players has completely changed. The so-called ‘universal’ has invariably fallen into the quagmire of customization, ultimately becoming a high labor consumption customized engineering service.
It’s not just startups that rely on exciting and grand stories to gain attention, even big companies are no exception. Upon closer examination of the financial reports of major companies, a common trend is to enhance the impact of AI on overall revenue and business growth. But often times, the market doesn’t pay. In the Hong Kong stock market, Baidu, Tencent, Ali and Kwai began to adjust their share prices to lower levels the day after the release of the financial report, and set new lows one after another.
In the final analysis, the original story telling model of the Internet is becoming more and more useless.
Death accelerates, AI companies should give up their illusions
If we observe the “death list” of AI companies, an obvious phenomenon is that the time for AI companies to go from being hot to being abandoned and forgotten is getting shorter and shorter.
For example, Sora will be released in February 2024, and the entire AI video industry will be beaten to death. All Internet giants will follow up in a hurry. It will take 25 months for Sora to go from stunning the world to “sudden death”. For example, British AI startup Robin AI was seen as a promising star in the European AI industry at the beginning of last year, completing multiple rounds of financing and attracting top global investors. In just six months, it was listed on the bankruptcy website.
AI has attracted the world’s most abundant funds, and they can quickly give birth to a unicorn. However, if the pace of this unicorn slows down, it may directly lead to elimination. There is no other reason, the evolution and turnover speed of the AI industry is too fast.
From text generation to image and video creation, from passive question answering to intelligent agent systems that actively execute tasks, the industry undergoes a paradigm shift every few months. Taking the general large model as an example, in 2024, the average iteration cycle of a large model version is about 132 days; By 2026, this number has been halved, and in some scenarios, even reduced to daily calculations. Therefore, under high pressure, the cycle of many AI startups from their high-profile debut to their dismal conclusion has been significantly shortened.
And a cruel fact is that the AI track seems to be blooming everywhere, accommodating numerous innovative companies, but the resources and technological advantages of the market are increasingly concentrated in a very small number of players, which also makes the survival situation of most AI companies not optimistic.
According to a report exclusively provided by private market data firm PitchBook to CNBC, nearly half of the 857 existing unicorn companies in the United States have not completed a new round of financing for over three years; More than 220 companies that once crossed the $1 billion valuation threshold have now been rebranded as’ fallen unicorns’.
Meanwhile, another set of numbers is overlapping in the same direction. Driven by the AI boom, capital has sent over $250 billion to OpenAI and Anthropic, and the number of global AI unicorns has skyrocketed from 245 at the end of 2024 to around 370 at the beginning of 2026. An extreme polarization pattern has emerged, with the top 10 most valued unicorn companies in the United States now accounting for 51.8% of the total market valuation, compared to only 18.5% in 2022.
Of course, even the few star companies currently standing at the top may not guarantee to retain their position in the next round of technological innovation.
So, in the future, this technological “competition” is likely to “kill” AI companies that are still trying to tell stories at the fastest speed, unlike in the past when they were given time to prove themselves or opportunities to find a home.
One sentence: When telling a story, one must be cautious.
相关文章
