As 2026 begins, the Hong Kong stock market is witnessing a concentrated wave of listings from "future stars" in the tech sector. KNOWLEDGE ATLAS and MiniMax (Xiyu Technology), the "AI Six Tigers" that emerged from the "Hundred-Model War," are set to list on the Hong Kong Exchange in quick succession. KNOWLEDGE ATLAS is scheduled to debut on January 8, becoming the "first stock for general AI large models," while MiniMax will follow a day later, planning to list on January 9.

Earlier, on January 2, the highly anticipated "tech star" BIREN TECH, one of the "Four Little Dragons of Domestic GPUs," successfully listed, claiming the title of "first GPU stock on the Hong Kong market."

All three companies were once considered potential star candidates for the A-share STAR Market, with BIREN TECH even initiating listing辅导 on the STAR Market before ultimately pivoting to Hong Kong. Why are high-quality tech companies collectively heading south? What considerations lie behind the capital? And what can they bring to the Hong Kong market?

For unprofitable tech companies, there are typically two listing paths: choosing the A-share STAR Market or listing on the Hong Kong Main Board via the Chapter 18C rules.

The STAR Market, launched in June 2019, offers five differentiated listing standards.

Regarding profitability requirements, only the first standard mandates profitability; standards two through five have no profitability threshold. The first standard requires a company to be profitable, with "positive net profit for the last two years and a cumulative amount ≥ RMB 50 million" or "positive net profit for the most recent year and revenue ≥ RMB 100 million."

In terms of market capitalization and revenue, all five standards incorporate market cap indicators, ranging from RMB 1 to 5 billion, while lowering revenue thresholds. The second standard, for instance, only requires revenue in the most recent year to be greater than or equal to RMB 100 million, a significant concession compared to the main board's requirement of ≥ RMB 1.5 billion.

Regarding cash flow, only the fourth standard requires that "the net cash flow from operating activities for the last three years totals ≥ RMB 100 million"; the other standards have no mandatory cash flow requirements.

As the "testing ground" for A-share registration-based reform, the STAR Market broke through traditional "profit-oriented" review requirements, pioneering an inclusive "market capitalization + multi-dimensional" indicator system. It also provides more market-friendly conveniences in review, governance, and refinancing, catering to the characteristics of tech companies having "long R&D cycles and high profit uncertainty."

The Hong Kong Exchange's Chapter 18C rules are similar to the A-share STAR Market requirements, serving as an important platform to support hard-tech company financing.

Chapter 18C, introduced in 2023, explicitly includes companies in AI, semiconductors, etc., abolishing sustained profitability requirements. In 2024, it further lowered market cap thresholds, reducing the requirement for Commercialized Companies from HK$6 billion to HK$4 billion, and for Non-Commercialized Companies from HK$10 billion to HK$8 billion.

Specifically, Chapter 18C lowers listing conditions across multiple dimensions, offering more flexible institutional arrangements.

Financially, Chapter 18C shows slightly more包容. For market cap, it requires HK$4 billion for Commercialized Companies and HK$8 billion for Non-Commercialized Companies.

Regarding profitability, Chapter 18C fully exempts profit requirements, allowing companies to list without profits or even with accumulated losses.

For revenue, Chapter 18C requires Commercialized Companies to have revenue ≥ HK$250 million in the most recent year, while Non-Commercialized Companies face no revenue requirement.

There are no mandatory cash flow requirements under 18C; it only requires Non-Commercialized Companies to have sufficient working capital to cover at least 125% of expenses for 12 months post-listing.

Regarding tech attributes, 18C pays more attention to R&D expenditure. It requires Commercialized Companies to have R&D costs ≥15% of operating expenditure, and Non-Commercialized Companies ≥30%. The STAR Market requires R&D investment ≥5%, or a cumulative amount ≥ RMB 80 million over the last three years.

In terms of equity structure, the flexibility advantage of HKEX's 18C is evident. It fully supports Red Chip enterprises listing without additional compliance obstacles and allows VIE-structured enterprises to list directly. It permits Weighted Voting Rights structures with no market cap threshold and no strict limit on the ratio of special voting rights, though it's typically not exceeding 10x. Both require a 3-year trading record, but 18C's requirements for operational continuity are more lenient.

The listing process under 18C also appears more efficient. It uses a dual-filing system with a review cycle of about 6-8 months, no mandatory辅导 period, and typically fewer inquiry rounds than the STAR Market. The STAR Market's registration-based process takes about 6-9 months, but加上 an average 8-month辅导 period, totals approximately 14-17 months. The public float requirement under 18C is ≥10%, with no mandatory new share issuance quantity; the STAR Market requires a public float ≥25%, reducible to ≥10% if share capital exceeds 400 million shares. Lock-up periods under 18C are 12-24 months for controlling shareholders, 12 months for core technical personnel, and 6 months for senior independent investors in Commercialized Companies (12 months for Non-Commercialized). The STAR Market requires a 36-month lock-up for controlling shareholders/actual controllers, 12 months for pre-IPO shareholders, with additional requirements for directors, supervisors, senior management, and core technical personnel.

The core simplifications of Chapter 18C lie in the fully exempted profit requirements, relaxed revenue restrictions, flexible support for Red Chip and WVR structures, shorter controlling shareholder lock-ups, and no mandatory辅导 period. The STAR Market, conversely, focuses more on industrial capability and financial stability, with relatively stricter multi-indicator requirements for tech attributes.

Applying this to companies like KNOWLEDGE ATLAS and MiniMax, do they meet STAR Market requirements?

KNOWLEDGE ATLAS reported revenue of RMB 191 million in the first half of 2025. Its IPO market cap exceeded HK$50 billion, approximately RMB 46.5 billion. Its 2024 R&D expenditure was >15% of operating支出, and it holds 84 invention patents, basically meeting the STAR Market's fifth standard.

MiniMax reported revenue of about RMB 376 million for the first three quarters of 2025, with an IPO market cap around RMB 36.4 billion. It maintains high R&D investment, with its R&D team comprising over 70% of staff. It also basically meets the fifth standard for unprofitable enterprises, though its Red Chip structure requires additional approval. Furthermore, the company has been operating for less than 3 years, not meeting the STAR Market's "3 consecutive years of operations" requirement. Using a VIE structure would require extra approval, potentially affecting the listing process.

Although both are now listing in Hong Kong via the 18C mechanism, there are slight differences. MiniMax's stock suffix includes "WP," where "W" denotes Weighted Voting Rights and "P" indicates failure to meet the commercialized revenue threshold. KNOWLEDGE ATLAS has no special suffix because its audited 2024 revenue reached HK$300 million, satisfying the HK$250 million commercialized revenue threshold, so it is not considered a "Non-Commercialized" enterprise.

Is it that "18C" is more attractive, or that Hong Kong stocks offer better "value for money"?

"While the A-share STAR Market supports tech companies, its implicit requirements for R&D ratio and revenue growth may deter unprofitable frontier tech companies," said Zhou Mingzi, Executive Director of Frost & Sullivan Greater China. For startups focused on AI large models, GPUs, and other frontier algorithms that are difficult to profit from in the short term, Hong Kong provides a financing outlet that is hard to replace in the mainland market.

Due to the capital-intensive nature of tech R&D, "high growth, no profit" is a common dilemma for KNOWLEDGE ATLAS, MiniMax, and BIREN TECH. Prospectus data shows that from 2022 to H1 2025, KNOWLEDGE ATLAS accumulated over RMB 4.4 billion in R&D投入. MiniMax reported an adjusted net loss of $186 million for the first three quarters of 2025. BIREN TECH accumulated over RMB 6.3 billion in losses over three and a half years, with a net loss still reaching RMB 1.601 billion in H1 2025. This profitability status may conflict with the STAR Market's "hidden thresholds."

A sponsor told SINA Corp, "The R&D成果 of AI large model and GPU companies are difficult to measure with traditional financial metrics, but the STAR Market's review logic emphasizes the certainty of profit prospects. These companies prioritize technological iteration over profitability, creating an inherent矛盾 between their development logic and listing rules."

BIREN TECH explicitly mentioned its reason for switching to Hong Kong in its prospectus: "The联交所 can provide the company with a platform to access capital and attract diversified investment," indirectly confirming the role of policy adaptability.

Furthermore, the choice of listing venue for tech companies involves an alignment between capital and corporate development needs, such as the compatibility of equity structures.

Prospectuses show that KNOWLEDGE ATLAS, MiniMax, and BIREN TECH all have foreign shareholders or special governance structures. Hong Kong is more包容 regarding Red Chip structures and dual-class share systems compared to the STAR Market.

MiniMax uses a WVR structure, where founder Yan Junjie holds approximately 28.25% of shares but controls about 74.91% of the voting rights. This design effectively safeguards founder control but conflicts with current A-share rules. KNOWLEDGE ATLAS's shareholder list includes not only domestic giants like Alibaba and Tencent but also international investment firms like Sequoia and Hillhouse, with cumulative pre-IPO financing exceeding RMB 8.3 billion, reflecting a diversified ownership structure. BIREN TECH's shareholders also include foreign institutions like PA GCC Limited and state-backed enterprises like Shanghai Lingang, and its governance structure without a controlling shareholder demands greater flexibility from listing rules.

"Currently, the A-share market cap threshold for Red Chip listings is relatively high. The Shenzhen Main Board requires a market cap of no less than RMB 200 billion for secondary listings of Red Chip companies. While the ChiNext board allows listings by Red Chip companies not listed overseas, its revenue and market cap requirements are out of reach for many startup tech firms," an investment banker pointed out. Hong Kong's attitude towards Red Chip and WVR structures is almost "fully accepting," allowing companies to list without complex equity restructurings, significantly reducing costs and time. Additionally, the "Tech Company Listing Channel" launched by HKEX in 2025 allows eligible companies to submit applications confidentially, which can effectively avoid premature exposure of trade secrets for companies like KNOWLEDGE ATLAS and MiniMax in critical R&D phases, helping them gain a competitive edge.

Advantages at the capital level further solidify the choice for a Hong Kong listing.

"The most important factor is the financing environment," an insider close to the aforementioned companies told SINA Corp. "The international capital in the Hong Kong market is the biggest attraction."

Li Jiacong, Director of Yisheng Fintech, told SINA Corp that in the Hong Kong market, cornerstone investors for many quality tech companies often include leading Chinese and international institutions. These cornerstones not only provide a stable subscription base for new issues but their professional endorsement also boosts market confidence. The willingness and feasibility of foreign institutions participating in Hong Kong IPOs are higher than in the A-share market, creating favorable conditions for companies to attract global long-term capital.

Furthermore, the Hong Kong market offers greater flexibility in capital operations. Li Jiacong stated that compared to A-shares, Hong Kong has simpler processes and more diverse methods for equity structure adjustments and refinancing. The average refinancing cycle in Hong Kong is 15 working days, allowing companies to quickly seize financing opportunities. For companies considering future actions like secondary financing or equity optimization, the Hong Kong market's efficient mechanisms provide support.

"For companies in fields like AI and GPUs, which require large R&D投入, long cycles, and high risks, the need for long-term, stable capital is more urgent. Hong Kong, as an international financial center, with its diversified investor base and global financing platform, can match the development needs of such companies," Li Jiacong said.

For companies with overseas business布局, a Hong Kong listing better aligns financing with business operations. MiniMax's overseas revenue accounted for 73.1% in the first three quarters of 2025. After listing in Hong Kong, the boost to its international profile will directly aid overseas market expansion. An analyst from Huaxin Securities pointed out that non-major-firm-backed companies like KNOWLEDGE ATLAS and MiniMax listing in Hong Kong not only secure financing for themselves but also provide valuation anchors for the entire AI industry, pushing the industry narrative from "technology stories" towards "commercial value realization."

However, it's noteworthy that the choice of listing venue is not exclusive; the path of "Hong Kong first, A-shares later" is a rational choice for companies balancing international financing with domestic capital market access. BIREN TECH stated in its prospectus that it will promote an A-share issuance上市 at an appropriate time, subject to compliance with rules.

After tech companies secure their "survival," how do IPO investors "survive"?

Liu Guoxian, Head of Capital Markets and Professional Practice at KPMG China, previously told SINA Corp that investing in unprofitable tech companies requires a certain 'threshold'. Investors need sufficient information to judge the business and commercial prospects of companies in high-tech-barrier fields like AI. These companies are high-risk and unstable, but that doesn't mean they are 'bad'."

For many unprofitable tech companies, listing in Hong Kong is more like a "battle for survival." "Many companies are rushing to list to fulfill VAM agreements and meet capital exit needs, rather than having a foundation for sustainable development," an industry practitioner from a consulting firm suggested.

Currently, market cooling signals are already apparent. The first-day break rate for Hong Kong IPOs rose to 29% in 2025, up from 21% the previous year. Among these, the break rate for unprofitable Specialised Technology Companies was as high as 30%. In November 2025, 5 out of 11 new listings broke on their first day, a break rate of 45.45%. On December 22, four new stocks, including Mingji Hospital and Huazhen Bio, collectively broke on their debut.

Former Hong Kong Financial Secretary John Tsang直言 he did not understand why four new stocks broke collectively, "Was some work not done properly? If we are not careful, Hong Kong's status as an international financial centre could be destroyed overnight." He also hinted that the prosperity of Hong Kong's IPO market背后 might be due to relaxed regulation and an "open door" policy, leading to aftereffects like IPO breaks and narrow shareholding concentration. "Although more funds can be absorbed now, the situation is like killing the goose that lays the golden eggs. Once the 'toy is broken,' funds leave, and only a mess remains."

Furthermore, several market analysts have stated that a large number of密集 new listings will dilute Hong Kong's market liquidity. Edmund Wong, Head of Listing Business, Deloitte China Capital Market Services Group, earlier told SINA Corp, "The trend of funds concentrating towards quality targets will become more pronounced." A sponsor revealed that密集 new listings lead to excessive dispersion of market funds, and small-to-mid-cap projects in Hong Kong are already experiencing undersubscription and被迫 scale reductions. Many mediocre unprofitable companies may face illiquid trading after listing. This differentiation could not only exacerbate valuation disorder but also affect the stability of Hong Kong's market ecosystem, weakening its overall appeal.

Even more alarming is the potential risk of a delisting wave. On December 29, 2025, HMVOD video received a delisting warning from HKEX due to sustained losses and asset shrinkage, becoming an缩影 of the operational difficulties faced by unprofitable companies post-listing. As the 3-5 year commercialization assessment period approaches, a large number of companies failing to meet profit targets may collectively face delisting reviews, potentially causing investor losses and triggering a market信任危机.

Regulators also face challenges. Observation reveals that Chapter 18C has "模糊界定" in its definition of "Specialised Technology," industry boundaries, calculation methods for R&D投入, recognition standards for companies not on the list, the division between Commercialized/Non-Commercialized, and special exclusion clauses for biotech companies.

For example, the indicative industry list provided by HKEX covers 5 major areas and 20 sub-sectors, including New Generation Information Technology, Advanced Manufacturing, Frontier Healthcare, New Materials, and New Energy. Sub-sectors include AI large model applications, semiconductors, cloud services (SaaS/PaaS/IaaS), advanced communications, robotics, biotechnology, new energy, and new energy vehicles.

However, it also explicitly states it "does not cover all acceptable industries and sectors." The current list update mechanism is opaque and lacks a fixed cycle. Emerging technologies like quantum computing and brain-computer interfaces are not included, and some交叉 fields like medical AI and industrial internet lack clear classification.

For companies not on the official list, HKEX has set four qualitative criteria.

First, high growth potential, but with no量化 indicators like CAGR or market size. Second, core business adopts new tech/business models, but the line between "new tech" and "traditional tech" is模糊, making it hard to distinguish technological from business model innovation. Cross-sector company classification is also problematic; if a firm involves both specialised tech and traditional business, there's no clear revenue占比 standard to judge if it's "primarily engaged" in specialised tech. Third, R&D contributes the majority of expected value, but "majority" lacks a specific ratio, making objective measurement difficult and reliant on subjective assessment. Fourth, R&D constitutes most of the expenditure, but "most" has no明确 threshold, potentially conflicting with the R&D投入 ratio requirement and creating double standards.

Additionally, 18C's tolerance for "non-specialised tech business" is unclear. It doesn't specify the maximum proportion of non-specialised tech business allowed, potentially enabling some companies to combine a small amount of specialised tech with大量 traditional business for listing. The exclusion criteria for "inter-segment revenue" are模糊, making it hard to judge if revenue from intra-group transactions should count towards specialised tech business revenue. There are also no clear restrictions on the types of non-specialised tech businesses, such as whether traditional sectors like real estate or finance are absolutely prohibited.

This policy "宽容" leaves room for pseudo-tech companies to exploit regulatory loopholes. Coupled with the difficulty in verifying the forward-looking, predictive information disclosed by unprofitable companies, less experienced investors are more susceptible to information asymmetry. These lower-quality companies,侥幸 listed, may see valuations inflated by speculative "dream multiples." Once market sentiment reverses, their stock prices plummet, often leaving retail investors as the ultimate victims.

Therefore, balancing leniency and strictness tests regulatory wisdom.

Currently, HKEX is building a market ecosystem that allows both entry and exit by improving delisting mechanisms and strengthening信息披露 requirements. Experts suggest introducing a dual "technical + commercial" assessor mechanism, involving industry experts in pre-hearing assessments to prevent excessive corporate packaging for listing. They also propose setting up liquidity buffers, mandating that a portion of IPO shares be allocated to long-term tech funds with lock-ups to stabilize market volatility risks.

The industry generally affirms HKEX's "milestone" move to support unprofitable tech listings, recognizing its importance for improving the sci-tech innovation financing system and promoting industrial upgrading. However, only through strict审核, sound regulation, rational investment, and corporate self-discipline can capital markets truly serve technological innovation. Whether the Hong Kong market can, through institutional improvement and ecosystem optimization, turn 18C into a genuine "tech incubator" rather than a "high-volatility minefield" will require further case studies for validation.