Preface: Unisound embarked on its journey to the STAR Market in November 2020 by submitting its prospectus to the Shanghai Stock Exchange, aiming to become the [First AI Speech Stock]. By 2023, the company redirected its efforts, submitting a prospectus for a main board listing to the Stock Exchange of Hong Kong Limited, with aspirations of achieving an IPO in Hong Kong's stock market. After five years of relentless pursuit, Unisound is poised to embark on a new chapter on the Hong Kong Stock Exchange.

Author | Fang Wensan
Image Source | Network

Hong Kong Welcomes Its First AGI Stock of the Year

According to the latest announcement on the official website of the Hong Kong Stock Exchange, Unisound Intelligent Technology Co., Ltd. has successfully navigated the listing hearing and released the post-hearing information set tonight.

CICC and Haitong International served as joint sponsors, collaboratively participating in the listing process.

Prior to this milestone, Unisound had completed the necessary filings with the China Securities Regulatory Commission regarding overseas issuance and listing, as well as the [full circulation] of unlisted shares within the territory.

This achievement signifies that Unisound is on the brink of becoming the first listed company with AGI as its primary business in the Hong Kong stock market this year.

Founded in 2012, Unisound is an AI enterprise dedicated to providing intelligent speech technology and comprehensive solutions.

Dr. Huang Wei, the visionary founder and CEO of the company, holds a degree from the University of Science and Technology of China and has a distinguished background, having worked at Motorola China Research Center and the Innovation Institute under Shanda Network.

In 2013, Unisound unveiled its first-generation full-stack AI system, equipped with perception and cognitive capabilities.

In 2017, the company introduced UniCore, its first large language model based on BERT. As the cornerstone algorithm model of Unisound Brain, this innovation offers a plethora of AI solutions tailored to various vertical industries.

In 2018, Unisound launched the UniOne series of edge interactive AI chips.

In 2023, the company unveiled Shanhai, a self-developed large model with a parameter scale of 60 billion.

In terms of computing power, Unisound boasts an Atlas AI intelligent computing cluster capable of over 184 PFLOPS and a storage capacity exceeding 10PB, providing robust support for technological research and development.

Furthermore, Unisound has supplied approximately 13 million consumer-grade AI chips to the market and facilitated the commercialization and application of automotive chips through NPU IP licensing.

The company also offers AI large language model capabilities to developers and enterprises on-demand through the MaaS model. Its product portfolio includes AI capability APIs based on public clouds, customized proprietary AI technology service platforms, AI model embedded chips, and IoT hardware modules.

Public records indicate that Unisound completed a total of 11 funding rounds prior to its IPO, amassing over US$340 million in cumulative financing. This impressive feat was achieved with the participation of over 30 investment institutions, including prominent names like Qiming Venture Partners, Trustbridge Partners, China Internet Investment Fund, JD.com Shangke, Qualcomm, and CICC.

According to Frost & Sullivan data, Unisound ranks fourth among AI solution providers in China in terms of revenue for 2024, with a market share of 0.6%. In the domestic life AI solution market, Unisound holds the third position, while in the realm of medical AI services and solutions, it ranks fourth.

According to the prospectus, the net proceeds from this IPO will primarily be allocated to investing in Atlas AI infrastructure to bolster R&D capabilities, exploring emerging business opportunities, enhancing product application and market penetration in vertical industries and scenarios, international expansion and strategic collaborations, as well as working capital and general corporate purposes.

Still Navigating Losses, Seeking Capital Market Support Becomes Crucial

Despite Unisound's revenue growth in smart life and smart healthcare, the company remains mired in losses.

From 2022 to 2024, Unisound's revenue amounted to RMB 600 million, RMB 730 million, and RMB 940 million, respectively.

Specifically, the smart life business generated revenue of RMB 490 million, RMB 580 million, and RMB 740 million, accounting for 81.0%, 79.6%, and 78.8% of total revenue during these years. Meanwhile, the medical business contributed RMB 110 million, RMB 150 million, and RMB 200 million, respectively, representing 18.9%, 20.4%, and 21.2% of total revenue.

Despite the annual revenue growth, the loss figures for the same period stood at RMB 375 million, RMB 376 million, and RMB 454 million, respectively, with the loss margin widening each year and showing no signs of reversal. Over the past three years, the company has accumulated losses exceeding RMB 1.2 billion and remains in the investment stage.

Currently, Unisound's scale is insufficient to address all facets, and it faces multiple disadvantages in terms of financial strength, technical accumulation, and market resources. This underscores the urgency for the company to seek support from the capital market.

Unisound is actively expanding into other application scenarios and vertical industries, presenting a broader commercialization horizon.

The company continues to delve deeply into artificial intelligence algorithms, achieving substantial results. This technological prowess serves as a cornerstone for its successful commercialization endeavors.

As of December 2024, the number of successfully filed large AI models at the national level has surged to 188, marking the intense onset of the [Hundred Model War].

Industry giants like Baidu, Tencent, and Alibaba are swiftly constructing their exclusive ecological large models, leveraging their vast business ecosystems.

However, in the fierce competition for large AI models, Unisound's investment scale pales in comparison to that of these industry leaders. For instance, the computing power of the Shanhai large model currently stands at 184 PFLOPS, significantly lower than Alibaba Cloud's 300+ PFLOPS and Baidu's 1840 PFLOPS.

Large models impose stringent requirements on algorithms, computing power, and data, which in turn necessitates a larger enterprise scale and stronger financial backing.

Conclusion:

Unisound's current predicament is not an isolated case but a microcosm of the AI industry bubble bursting.

Industry data reveals that among the 188 filed large models in China in 2024, 90% have annual revenues of less than RMB 10 million. In the primary market, the VC/PE financing amount in the AI sector decreased by 47% year-on-year in 2024, indicating a shift in capital's focus from [stories] to [profits].

Some references: Taimedia AGI: "This year's [First AGI Stock in Hong Kong Stocks] has been confirmed. Unisound has finally passed the Hong Kong Stock Exchange hearing after five years of rushing for an IPO", Meicai Network: "After a loss of RMB 1.2 billion in three years, Unisound rushes for an IPO again, trapped in the AI dilemma", Yiou Network: "[Unisound] passed the Hong Kong Stock Exchange hearing. CEO Huang Wei said, '[Artificial intelligence is my lifelong career]'", Digital Intelligence Research Society: "Unisound, which was [exposed as fake], has suffered losses of over RMB 1.2 billion in three years"