Peter Drucker said that the best way to predict the future is to create it.

In the context of the Hong Kong stock market's AI IPO market, this statement has a more realistic annotation: capital is certainly willing to pay for the future, but the premise is that the future has already begun to enter revenue, orders, gross margins, and customer structures.

On July 8, the Hong Kong Stock Exchange welcomed five companies to list on the same day, among which Momenta, RECONOVA, and EACON all carry distinct AI labels, corresponding to autonomous driving software, visual AI, and unmanned driving in mining areas, respectively.

Among them, Momenta had an issue price of HKD 295.6, opened at HKD 301, and raised approximately HKD 5.89 billion; RECONOVA saw a significant drop on its first day; EACON gained relatively stable support due to the certainty of unmanned driving scenarios in mining areas.

Zooming out, in the first half of 2026, Hong Kong's IPO fundraising reached HKD 210 billion, a year-on-year increase of 92%, with 87 listings, setting a new high for the same period in nearly five years. The wave of AI listings in Hong Kong has not stopped; the capital market is simply re-differentiating: which companies are still creating the future, and which companies have already begun to realize it.

The AI listing wave continues, but funds are beginning to separate the two words "AI" for re-pricing

The Hong Kong market remains an important window for global tech company financing in 2026. According to PwC, the IPO fundraising amount in Hong Kong reached HKD 210 billion in the first half of this year, ranking second globally; among the main board new stocks, information technology and telecommunications services accounted for 48%, industrial and materials accounted for 21%, and healthcare and pharmaceuticals accounted for 18%. The Hong Kong Stock Exchange disclosed that as of the end of April 2026, among the 15 companies listed under 18C, 14 came from different segments of the AI value chain. The supply of AI assets has not decreased, and institutions, funds, and industrial companies are still concentrating on the Hong Kong stock market.

Changes have occurred in the way buyers filter. From 2024 to 2025, as long as AI companies have labels such as large models, autonomous driving, robotics, chips, and visual intelligence, the market is often willing to give a higher imagination space. In 2026, funds pay more attention to business penetration: whether revenue comes from one-time projects or replicable software licenses; whether growth relies on government or large customer budgets, or comes from real commercial scenario expansion; whether gross margin improvement comes from economies of scale or short-term project structural fluctuations.

The simultaneous listing of Momenta, RECONOVA, and EACON provides a sample. Momenta represents the broadest story of intelligent driving platforms, with clients including Mercedes-Benz, Toyota, BYD, SAIC, Audi, and other automakers; RECONOVA represents the industry landing of visual AI in civil aviation, commercial spaces, and safe driving; EACON represents the transition of unmanned driving in closed mining areas from pilot projects to large-scale deployment. All three companies can be categorized as AI assets, but their revenue quality, customer structure, investment intensity, and profit paths are completely different For AI new stocks, the performance on the first day of listing is not just about the rise and fall of one day, but rather the feedback from the secondary market to the industry: capital no longer prices uniformly based on "AI concentration," but requires companies to convert technological advantages into clearer revenue models and cash flow paths.

Momenta focuses on software licensing, RECONOVA focuses on project quality, and EACON focuses on mining scene replication.

Momenta's business vision is the most complete and the heaviest. The prospectus shows that the company's revenue from 2023 to 2025 is expected to be 743 million yuan, 1.325 billion yuan, and 2.413 billion yuan, respectively; during the same period, the annual losses are expected to be 2.57 billion yuan, 3.206 billion yuan, and 3.458 billion yuan. The company's revenue growth is rapid, but R&D investment and the fair value changes of financial liabilities such as preferred shares still put pressure on the profit and loss statement. By 2025, Momenta's R&D expenditure will reach 1.869 billion yuan, with the proportion of revenue decreasing from 113.9% in 2024 to 77.5%, indicating that economies of scale have begun to appear, but absolute investment remains high.

The most important variable for Momenta is not "whether it can talk about autonomous driving," but whether the mass production vehicle solution can continuously increase the proportion of licensing fees. The prospectus reveals that the revenue model for the company's mass production vehicle solutions includes technical development service fees before mass production and licensing fees after mass production, with licensing fees collected as a one-time fee per equipped vehicle; the proportion of licensing fees in total revenue has increased from 3.1% in 2023 to 22.1% in 2024, and further to 40.1% in 2025. This set of data is more critical than the Robotaxi story because software licensing revenue is closer to standardization and easier to release profit elasticity.

The challenges for RECONOVA lie on the other end. The company's revenue from 2023 to 2025 is expected to be 242 million yuan, 395 million yuan, and 443 million yuan, with gross margins of 40.8%, 39.8%, and 37.7%, respectively; the net loss in 2025 is expected to be 68.149 million yuan. Revenue growth is still present, but the growth rate in 2025 is significantly slowing down, with a slight decline in gross margin and net profit turning negative again. The business structure also indicates issues: in 2025, smart civil aviation revenue is expected to be 172 million yuan, accounting for 38.9%; smart commercial revenue is expected to be 154 million yuan, accounting for 34.9%; smart safe driving revenue is expected to be 116 million yuan, accounting for 26.2%. The application scenarios for visual AI are numerous, but having multiple scenarios does not inherently equate to high-quality growth.

RECONOVA resembles a company caught between "AI productization" and "project-based delivery." Civil aviation, parks, and freight safety all have budget entry points and industry barriers, but project rhythm, acceptance cycles, customer concentration, and accounts receivable pressure can all affect profit quality. Munger once said that all smart investments belong to value investing, and the key is to estimate value. For RECONOVA, the market needs to estimate not only the visual AI capabilities but also whether project delivery can be transformed into reusable products and whether smart commercial and safe driving can create new growth curves beyond civil aviation The story of EACON is not as grand as that of open-road autonomous driving, but its financial and operational metrics are easier to understand. The prospectus shows that the company's revenue from 2023 to 2025 is projected to be 271 million yuan, 986 million yuan, and 1.435 billion yuan, respectively, with the revenue from autonomous mining truck products and solutions in closed environments reaching 1.428 billion yuan in 2025, accounting for 99.5% of total revenue. The gross margin is expected to improve from -18.6% in 2023 to 7.6% in 2024, and further increase to 10.1% in 2025. By the end of 2025, the number of active autonomous mining trucks is expected to reach 2,580, deployed across 42 mining sites.

The advantage of autonomous driving in mining areas lies in the more direct demand. Open-pit mining roads are relatively closed, with fixed transportation routes, allowing for clear ROI in terms of safe production, labor substitution, and cost reduction and efficiency improvement. Based on the number of active autonomous vehicles by the end of 2025, EACON's market share is expected to reach 55.5%, ranking first in China's autonomous mining solution market; in terms of revenue, the company is projected to generate approximately 1.4 billion yuan in 2025 in this market, with a market share of 37.6%. This business does not rely on long-term imagination to support its entire value; orders, fleet size, number of mining sites, and gross margin are the core verification metrics.

The next round of competition will focus on revenue structure and profit quality.

The AI IPOs in the Hong Kong stock market have not entered a cold phase but rather a stratification. The key for MOMENTA is whether the proportion of mass production software licensing revenue can continue to increase, while the Robotaxi business should not excessively consume cash flow; for RECONOVA, the key is whether visual AI can transition from project revenue to higher reuse product revenue; for EACON, the key is whether the replication of mining scenarios can drive further improvement in gross margin while reducing the capital occupation of heavy asset fleet models.

The competition for AI companies after going public will no longer stay at the level of technical terminology. The capital market will look at how much customer coverage is gained from R&D expenses, whether customer coverage can turn into orders, whether orders can convert into revenue, and whether revenue can lead to improvements in gross margin. While the technical narrative remains important, it must be confirmed through the income statement.

The simultaneous listing of the three companies provides a clear signal to the Hong Kong stock market: AI assets remain a mid-term mainline, but the fundamental differences among different AI companies will be amplified more quickly. Autonomous driving, visual AI, and autonomous mining are all seeking positions in industrial substitution, and the companies that ultimately succeed may not necessarily be the ones with the biggest stories, but rather those that first integrate AI capabilities into customer budgets, delivery cycles, and cash collections.

This also marks the true maturation of the market. The market has not abandoned AI; it has simply begun to demand that AI companies answer a harder question: after technological leadership, where does profit come from?