NewTimeSpace | Continued Losses Exceeding 900 Million RMB! "Oversubscription King" Deepexi Technology (01384.HK) Market Cap Shrinks Nearly 16 Billion HKD in 5 Trading Days

As the 2025 "Oversubscription King" under the HKEX Chapter 18C rule, Deepexi Technology (01384.HK) has recently experienced severe share price volatility. As of March 12, 2026, it closed at HK$69.80, with a single-day drop of 15.60% and a cumulative crash of nearly 40% over five trading days, reducing its market cap to HK$22.8 billion. While revenue is strong, losses persist; the company expects 2025 revenue of RMB 401M to 425M (up 65%-75% YoY), with core FastAGI growth exceeding 175%. However, estimated net losses remain as high as RMB 905M to 965M, bringing cumulative losses over the last four years to over RMB 3 billion.

The 2025 "Oversubscription King" of the Hong Kong stock market, Deepexi Technology (01384.HK), which listed under the Chapter 18C regime, has seen its share price continue to tumble recently.

Data shows that as of the close on March 12, 2026, the company was quoted at HK$69.80, with a single-day drop of 15.60%. Its market capitalization has shriveled to HK$22.8 billion, representing a wipeout of nearly HK$16 billion in value compared to March 5.

Persistent Losses Hinder Profitability; Nearly 40% Decline in 5 Trading Days

March 6, 2026, marked the turning point where Deepexi’s share price shifted from an uptrend to a sharp decline.

NewTimeSpace understands that on this day, Deepexi issued a positive profit alert. The company expects revenue for the year ending December 31, 2025, to range between approximately RMB 401 million and RMB 425 million, representing a year-on-year increase of roughly 65% to 75%.

The company noted that revenue growth was primarily driven by a significant increase in income from its FastAGI enterprise-level AI solutions, which is expected to grow by over 175% compared to 2024.

However, the announcement also estimated a net loss for the reporting period between approximately RMB 905 million and RMB 965 million. While this represents a narrowing of more than 23% compared to 2024, the figures remain stark.

Notably, despite the sustained surge in revenue, the company remains deep in the red. On the day of the performance announcement, Deepexi’s shares fell by 3.90%, and the downward momentum accelerated following the full profit forecast.

As of the close on March 12, 2026, Deepexi was priced at HK$69.80, with a single-day plunge of 15.60% and a cumulative decline of 39.04% over the last five trading days.

NewTimeSpace has learned that Deepexi has been loss-making for several years. Data shows that from 2022 to 2024, revenues were RMB 100.5 million, RMB 129 million, and RMB 243 million, respectively, showing rapid top-line growth.

Conversely, losses expanded significantly during the same period, with annual net losses recorded at RMB 655 million, RMB 503 million, and RMB 1.255 billion, respectively.

The continuous losses suggest that revenue growth is being "swallowed" by high operating costs. Although the 2025 loss is expected to narrow compared to 2024, it remains high relative to 2022 and 2023. Cumulative losses over the past four years have exceeded RMB 3 billion.

"China's Palantir" Supported by Two Core Pillars

Public records show that Deepexi was founded in 2018. Its primary business focuses on enterprise-level Large Language Model (LLM) AI application solutions. Its technological foundation rests on two core platforms: the FastData Foil data fusion platform and the Deepexi Enterprise LLM Platform.

NewTimeSpace understands that the company adopts a project-based service model. Its core business consists of two main sectors: FastData enterprise-level data intelligence solutions and FastAGI enterprise-level AI solutions.

Currently, these two solutions have achieved large-scale implementation across several vertical industries, serving a cumulative total of 283 enterprise customers in sectors such as consumer retail, manufacturing, healthcare, and transportation. The business focus is gradually shifting toward FastAGI.

As a provider of enterprise-level LLM solutions, the market must bear high computational costs; deploying large models requires massive computing resources, including high-end GPUs and significant energy consumption.

It is widely noted that Deepexi exhibits high similarity to Palantir in terms of technical architecture and business model. Both companies focus on "Data Governance + Intelligent Decision-making" as their core to provide end-to-end AI solutions for enterprise clients.

Deepexi has been dubbed "China's Palantir" due to its precise alignment with Palantir’s technical path. The latter is a US stock giant that saw its share price surge 26 times in three years, with a market cap exceeding US$400 billion.

During its IPO placement, Deepexi was fervently pursued by the capital markets. Its public offering was oversubscribed by 7,569.83 times, making it not only the "Oversubscription King" of 2025 HKEX Main Board IPOs but also ranking first in Hong Kong's history for oversubscription multiples on the Main Board.

On October 28, 2025, Deepexi listed on the Main Board of the HKEX, opening with a 111.93% surge and finishing its first day significantly above its offer price of HK$26.66.

Prior to this recent continuous decline, as of March 5, 2026, the company had achieved a cumulative gain of 342.61%. Following the recent sell-off, that gain has narrowed to 161.82%.

Industry Enters Explosive Phase; Market Competition Intensifies

With the promotion of policies like the "AI+" initiative and an explosion in industrial demand, the domestic enterprise-level LLM AI application track is entering a golden era.

Deepexi’s performance explosion is not an isolated case but a microcosm of the industry’s growth phase. According to Frost & Sullivan, the market size is expected to grow from RMB 8.4 billion in 2025 to RMB 52.7 billion by 2029, representing a CAGR of 58%.

NewTimeSpace has learned that Deepexi faces fierce market competition. Tech giants like Baidu and Alibaba are leveraging their general-purpose LLM technology to penetrate downward. For instance, Baidu AI Cloud’s Qianfan platform provides model training and application development tools, while Alibaba Cloud integrates AI capabilities into its data mid-end solutions. Additionally, firms like SenseTime and Megvii are positioning themselves in niche scenarios like AI vision, creating differentiated competition.

In Deepexi’s specific niche, 4Paradigm focuses on Decision AI, achieving large-scale implementation in finance and manufacturing via a "Model-as-a-Service" model, standing alongside Deepexi at the forefront of the track.

The Frost & Sullivan report shows that by revenue in 2024, Deepexi ranked fifth in China's enterprise-level LLM AI application solution market with a 4.2% market share. Meanwhile, the market shares of Baidu, iFlytek, Alibaba, and SenseTime were 11%, 9.7%, 7.3%, and 6.9%, respectively.

At present, although Deepexi’s losses have narrowed significantly, it is still far from a true profitability inflection point. As traditional software giants, internet majors, and AI startups all enter the fray, industry competition may intensify price pressure. Amidst this influx, the company may face pressure in customer acquisition. Whether it can continue to expand its large enterprise client base and increase ARPU (Average Revenue Per User) remains to be seen.

Based on 2025 performance, Deepexi’s Price-to-Sales (P/S) ratio was at a high level before the decline, suggesting that growth expectations were already fully priced in. If subsequent performance fails to meet expectations, the pressure for a valuation correction will be significant.

NewTimeSpace Disclaimer: All content herein is the original work of NewTimeSpace. Any reproduction, reprinting, or use of this content in any other manner must clearly indicate the source as "NewTimeSpace". NewTimeSpace and its authorized third-party information providers strive to ensure the accuracy and reliability of the data, but do not guarantee the absolute correctness thereof. This content is for reference only and does not constitute any investment advice. All transaction risks shall be borne by the user.

×
Share to WeChat

Open WeChat, use the "Scan", and share to my Moments.