The listing of WENGE AI (HKEX: 01956) on the Hong Kong Stock Exchange in June 2026 marks another AI company accessing the capital markets this year. Amid the global generative AI wave, persistent losses are no longer the sole metric for valuing AI firms, with industry focus shifting to technological moats, market positioning, growth potential, and paths to profitability. As a decision intelligence leader backed by the Chinese Academy of Sciences (CAS), WENGE AI was once viewed as a premium candidate within the national AI initiative. However, its prospectus reveals multiple concerns beneath the surface: revenue growth lagging the broader industry, profit improvement reliant on non-recurring items, volatile R&D investment, and several risks associated with the offering itself. While the market enthusiasm for new AI listings is high, WENGE AI's path to listing reflects both sectoral tailwinds and underlying business vulnerabilities.

Strong Pedigree and State-Backed Capital Support

WENGE AI's origins and development are deeply rooted in the CAS ecosystem. Founded in 2017 by AI scientists from the Institute of Automation, Chinese Academy of Sciences, its core team comprises over 30 senior AI scientists. The three co-founders are all industry experts in artificial intelligence and decision science. Over 26% of the R&D team hold master's or doctoral degrees, with an average of over eight years of industry experience. By the end of 2025, the company held 154 registered patents and 439 software copyrights, and co-owned several core patents with research institutions like the CAS Institute of Automation and the China Academy of Chinese Medical Sciences, forming the cornerstone of its technological development.

This pedigree is reflected not only in its technical foundation but also in its impressive investor roster. Prior to the IPO, the company completed ten funding rounds, attracting a star-studded list of investors including national-level funds and prominent institutions such as the China Development Bank Manufacturing Transformation and Upgrading Fund, the China Internet Investment Fund, and Shenzhen Capital Group. This "all-star" lineup provided solid support for initial market confidence.

Narrowing Losses Driven by Sales Cost Reduction; Volatile R&D Spending Raises Concerns

However, pedigree does not directly translate to robust financial performance. A closer examination of the financial data in the prospectus reveals that the logic behind WENGE AI's profit improvement is not as straightforward as it seems, with several areas of concern emerging.

Revenue growth, while steady, has failed to outpace the industry. According to the prospectus, revenues for 2023, 2024, and 2025 were RMB 250 million, RMB 320 million, and RMB 410 million, respectively, representing year-on-year growth rates of 27.3% and 27.6%. Although notable, this pace lags significantly behind the 44.4% growth rate estimated for China's large-model-driven decision intelligence market in 2025, suggesting the company's market share expansion may be slower than the industry average.

A notable divergence exists between the modest improvement in gross margin and the significant narrowing of net losses. From 2023 to 2025, the company's comprehensive gross margin increased only slightly from 44.0% to 51.2%, a cumulative improvement of less than 7 percentage points over three years, with a mere 0.8 percentage point increase from 2024 to 2025. In contrast, the net loss ratio shrank dramatically from 104.0% to 41.0% over the same period, with an 8.4 percentage point improvement from 2024 to 2025. This change was not driven by improved core profitability but by two non-core factors.

Firstly, sales and marketing expenses were sharply reduced from RMB 120 million in 2023 to RMB 94 million in 2025, with the ratio to revenue plummeting from 47.3% to 23.2%. Secondly, government grants and subsidies reached RMB 45 million in 2025, nearly doubling year-on-year. The sustainability of this "loss reduction," achieved through cost compression and external non-recurring income, is questionable.

More concerning is that, despite various measures to reduce losses, the net loss for 2025 still amounted to RMB 170 million, showing a slight increase from the RMB 160 million loss in 2024. The fundamental issue of "revenue growth without profit growth" remains unresolved. The company also explicitly warned in its prospectus that it expects to continue incurring losses in 2026.

Secondly, volatility in core R&D investment casts a shadow over the company's technological leadership. R&D expenses were RMB 180 million in 2023, dropped significantly to RMB 130 million in 2024—a notable decline in both absolute terms and as a percentage of revenue—before recovering to RMB 190 million in 2025. This periodic contraction in R&D spending not only affects the pace of technological iteration but also makes it challenging to maintain a long-term technological moat in the fiercely competitive AI sector.

Furthermore, issues with operating cash flow and receivables highlight operational pressures. From 2023 to 2025, net cash outflow from operating activities was RMB 180 million, RMB 130 million, and RMB 190 million, respectively. This indicates that despite booking revenue growth, the company has not generated positive cash flow, making its daily operations heavily reliant on financing.

Simultaneously, the trade receivables and notes turnover days climbed from 148 days in 2023 to 200 days in 2025, while the receivables balance increased from RMB 120 million to RMB 250 million. Extended collection cycles and a growing receivables balance not only tie up working capital but also introduce potential bad debt risks, further exacerbating cash flow pressure.

Business Model Flaws: Customization Hinders Scalability; Insufficient Competitive Barriers

Beyond financial concerns, WENGE AI's business model and competitive landscape present significant challenges.

In terms of business structure, on-premise deployment constituted the largest segment, accounting for 72.7% of revenue in 2025. While this project-based, customized service model meets specific client needs, it suffers from inherent weaknesses: long development cycles, high labor costs, and poor replicability, making scalable business expansion difficult.

Moreover, the company's customer retention rates for 2023, 2024, and 2025 were 55.1%, 65.4%, and 55.4%, respectively. According to the prospectus definition—clients from the previous year who remain clients in the current year—this indicates that nearly half of the existing customer base was lost in each of these three years.

Regarding industry competition, although WENGE AI ranks first in the decision intelligence niche, it holds only eighth position in the broader enterprise large-model market, with a mere 2.2% market share, suggesting its competitive edge as a sector leader is not pronounced. This segment is contested by major AI players like ByteDance, Alibaba, Baidu, and SenseTime, indicating that WENGE AI's technological and market share moat is not robust.

Multiple Uncertainties Surround the Hong Kong IPO Offering

The cornerstone investor subscription for this IPO totals only 27% of the offering, a relatively low proportion for Hong Kong new listings. The participating entities are primarily domestic institutions, with a notable absence of major international long-term capital. This limits the volume of funds committed to providing a stable, locked-up support base post-listing. The inclusion of Livermore Securities in the underwriting syndicate, a brokerage known for its online retail investor focus, has raised market questions about the level of genuine professional institutional endorsement, increasing concerns over post-listing share price volatility.

Valuation contradictions are even more striking. The company's post-money valuation after its last pre-IPO funding round was RMB 3.62 billion. The implied market capitalization at the IPO price is approximately HK$10.5 billion (around RMB 9.63 billion), representing an increase of nearly 166%, which significantly outpaces the company's 27.6% revenue CAGR.

Additionally, the price-to-sales (P/S) ratio for this offering is as high as 22.6x, second only to peer DeepTech among comparable companies. Reviewing recent trends for AI IPOs in Hong Kong reveals that thematic hype often supports only short-term price surges, with medium-to-long-term performance entirely dependent on fundamental differentiation: while shares of companies like Zhipu AI and MiniMax have maintained high valuation sentiment post-listing, others like Minglue Tech and DeepTech have experienced significant corrections from their post-IPO highs.

Prospectus Discrepancies Raise Questions on Internal Controls and Disclosure Rigor

The prospectus contains two basic data errors in the historical share transfer section: one transaction incorrectly states the transfer amount, and another involves a mistake in converting between millions and thousands of yuan. Such professional oversights have led the market to question the internal controls and disclosure rigor of both the company and its intermediaries.

Furthermore, the company completed capital reserve transfers to share capital on a pro-rata basis in 2019 and 2022, proportionally increasing all investors' shareholdings. However, the prospectus only discloses the original subscription prices for each funding round without adjusting and disclosing the diluted cost per share after these two transfers. This lack of transparency makes it difficult for the market to directly compare historical financing valuations and understand the equity evolution.

For investors, the listing of WENGE AI resembles a gamble between AI sector euphoria and the company's underlying fundamentals. Short-term market fervor may temporarily obscure operational weaknesses, but as sentiment cools, the company's true value will inevitably be reassessed. In an era crowded with AI listings, only companies possessing genuine core technological advantages, sustainable growth capabilities, and clear paths to profitability can establish a firm footing in the long-term capital markets. WENGE AI evidently has more to prove regarding its long-term value proposition.