Following the listing of Changxin Technology, August 19 marked another milestone as Unitree Technology surged 629.44% at its opening on the STAR Market, becoming the first "humanoid robot stock" on the A-share market. The consecutive debuts of these two flagship companies have ignited capital markets and signaled a powerful shift: China's capital markets have entered a golden era of "value reassessment" for hard-tech enterprises.

Beyond the spotlight, a diverse array of capital sources—including bank financial asset investment companies (AICs), wealth management firms, insurers, brokerages, state-guided funds, and internet CVCs—are rushing with unprecedented enthusiasm to back the hard-tech future. Yet, even within the same hard-tech arena, banking-affiliated capital, represented by AICs and wealth management funds, has adopted markedly different approaches toward the two benchmark companies, Changxin Technology and Unitree Technology. Upon closer inspection, this divergence is not a matter of favoritism but a rational division of labor driven by industry stage, capital attributes, and risk appetite.

On August 19, Hangzhou-based robotics newcomer Unitree Technology successfully listed on the STAR Market, officially making its debut as the A-share market's first humanoid robot pure-play. In early trading, the stock opened at 1,100 yuan per share, surging 629.44% from its issue price and pushing its market cap to 444.9 billion yuan. By the close, Unitree settled at 845 yuan, up 460.34% from the IPO price, with a market valuation of 341.8 billion yuan.

Not long before, Changxin Technology's listing also drew widespread attention. As two flagship hard-tech IPOs in the AI landscape, Changxin focuses on storage and computing power, advancing the self-reliance of domestic chips, while Unitree uses humanoid robots to bring intelligent algorithms from the virtual world into real-world applications. Together, they paint a clear picture of AI's journey from technological breakthroughs to industrial deployment. Behind them, a vibrant scene of financial capital eagerly "betting" on the hard-tech future is visible.

However, not all funds are moving in lockstep within this wave of capital aggregation. Even among banking-affiliated capital, AICs and wealth management funds have shown distinct preferences and strategies when faced with these two hard-tech benchmarks. For Changxin Technology, a leader in storage chips, its clear business model, mature technology roadmap, and well-defined import substitution logic have made it a prime target for heavy bank-backed investment. On the equity front, all five major state-owned bank AICs have taken stakes. The prospectus shows Agricultural Bank of China's AIC (ABC International) holds 0.95%, making it the largest bank-affiliated AIC investor; CCB International holds 0.83%, and with indirect holdings through CCB International and CCB Leading, the China Construction Bank system's combined stake reaches approximately 1.7%, ranking first among the five major banks. ICBC International holds roughly 0.64% through its ICBC Rongjin Fund, while Bank of Communications and Bank of China's AIC subsidiaries each hold 0.38%. Additionally, joint-stock and city commercial banks, including China Merchants Bank, Shanghai Pudong Development Bank, and Huishang Bank, have indirectly locked in exposure to the storage chip sector through private equity fund participation.

Wealth management firms have also been active during the IPO phase. According to Changxin's offline placement data, 29 wealth management products from institutions such as Minsheng Wealth, Bank of Nanjing Wealth, Bank of Ningbo Wealth, Industrial Bank Wealth, and Postal Savings Bank of China Wealth successfully secured 4.544 million shares, totaling 39.3514 million yuan in allocated value. In contrast, Unitree Technology's pre-IPO shareholder list shows no direct AIC equity investment; only wealth management firms participated in IPO subscriptions. Data reveals that six wealth management companies made it onto Unitree's offline placement list, with Everbright Wealth and Bank of Ningbo Wealth as the primary participants, each deploying dozens of products. China Merchants Bank Wealth, Postal Savings Bank of China Wealth, Minsheng Wealth, and Bank of Nanjing Wealth followed with four, two, one, and one product(s), respectively.

Why would banking-affiliated capital take such divergent paths toward two equally sought-after hard-tech targets? The answer lies in the fundamentally different evaluation frameworks of AICs and wealth management firms. Xue Hongyan, a special researcher at Subo Bank, explains that AICs adopt an industry-strategic perspective, prioritizing alignment with national strategy, irreplaceability in the supply chain, long-term technological evolution capability, potential for investment-loan linkage, and certainty of exit paths, with high tolerance for short-term profitability. In contrast, bank wealth management leans toward a financial and market-trading perspective, focusing on valuation reasonableness, short-term earnings delivery, post-listing liquidity, and net value volatility, with a stronger emphasis on absolute returns and net value stability.

In terms of capital attributes, AICs primarily deploy mid-to-long-term capital, capable of withstanding extended cycles and favoring large-scale targets with clear industry positioning. Wealth management funds, with shorter durations and lower risk appetite, mainly engage in IPO subscriptions and private placements within public market channels, treating hard-tech as an earnings enhancement within "fixed income plus" strategies. "The boundaries are clear," says Zeng Gang, president of Tianfu Liyan Financial Research Institute. "AICs assume the role of risk capital during the industrial incubation phase, while wealth management funds act more as liquidity providers and 'buyers' in the capital market, rarely delving into early-stage high-risk equity positions."

Regarding why AICs are absent from Unitree's shareholder list, some market views attribute it to a timing mismatch. A look at the timeline shows Unitree has been highly sought after by capital since its inception, with a tight early-stage funding cadence. Meanwhile, AICs under the five major state-owned banks were only established from 2017 onward, with equity investment pilots gradually expanding until March 2025, when the pilot scope was further widened, allowing AICs to enter a phase of intensive deployment. Wang Pengbo, chief analyst at Botong Consulting, notes that the maturation of AIC investment systems coincided with a period that fell outside Unitree's core financing cycle, compounded by geographic restrictions in early AIC equity pilots, which would have narrowed project search scopes.

Pan Helin, a renowned economist and member of the Ministry of Industry and Information Technology's expert committee on information and communication economics, offers a different take. He suggests that AICs, being return-oriented, were influenced by the cautionary tale of Boston Dynamics' struggles, and in Unitree's early days, the company had not yet demonstrated standout investment value. Unitree only gained widespread market attention starting with the 2025 Spring Festival Gala, so in its early phase, AICs may not have even been aware of the company, making this less about timing and more about visibility. He adds that investing in the tech sector inherently carries uncertainty, but AICs, with their information channel advantages, are well-positioned to engage at relatively earlier stages compared to ordinary investors.

"The time window does have an impact, but it's not the core factor," Xue Hongyan adds. The deeper reason, he argues, lies in the nature of the two sectors. Changxin's storage chip industry has a clear technical path, large asset scale, and high commercialization certainty, fitting the AIC preference for heavy assets, high certainty, and capacity to absorb large, long-term capital. Unitree's humanoid robotics sector, by contrast, is still in its industrial infancy, with rapid technological iteration and high valuation volatility, which deviates from AIC investment preferences. Moreover, later-stage funding rounds have been largely claimed by top-tier venture capital and industrial capital, further narrowing AIC participation opportunities.

"As bank-affiliated capital, AICs' risk appetite and assessment mechanisms push them to avoid early-stage, high-uncertainty tracks. The divergence reflects both objective timing constraints and an institutional preference for 'heavy manufacturing and high certainty' in sector selection," Zeng Gang states bluntly. Looking ahead to AICs' strategic investment logic in primary markets, Zeng predicts they will continue to follow the banking system's inherent characteristics, emphasizing cash flow predictability, verifiable technological moats, and clear exit paths, unlike pure market-oriented venture capital that prizes sector imagination and team dynamism. In their evaluation framework, AICs will focus on order visibility, patents, supply chain positioning, and balance sheet structure, while adding differentiated metrics such as alignment with national strategic industry catalogs, local government industrial synergy, and the ability to leverage the parent bank's existing client resources.

For future investment priorities, Zeng anticipates that hard-tech sectors with strong commercialization certainty and urgent import substitution needs—such as semiconductor equipment and materials, industrial mother machines, new energy storage, and biopharmaceutical pilot platforms—are more likely to attract AIC interest. Conversely, frontier sectors still in the exploratory phase of business models are unlikely to enter their core allocation range in the near term.

In contrast to AICs' "heavy stake in Changxin, absence in Unitree," wealth management firms have also shown differentiated participation patterns across the two projects, though in different directions. For instance, institutions like Everbright Wealth and China Merchants Bank Wealth participated only in Unitree's IPO subscription and did not engage in Changxin's offline placement. Xue Hongyan attributes this to differences in target characteristics and internal evaluation frameworks. Unitree's scarcity value and sector sentiment heat, combined with a small free float, align well with the short-term return goals of IPO subscriptions. Changxin, on the other hand, has a larger scale and cyclical attributes, with relatively weaker stock price elasticity, prompting some institutions to exercise greater caution.

A spokesperson from Bank of Ningbo Wealth also notes, "Humanoid robots and storage chips are at completely different industry cycles, requiring distinct assessment focuses. Humanoid robots are still in the 0-to-1 industrial introduction phase, with commercial scenarios not yet fully proven. Storage chips, meanwhile, represent a classic cyclical industry that is currently benefiting from AI-driven demand, placing the sector on an upswing. Our investment research hinges on accurately gauging the direction of the industry cycle."

Looking forward, Bank of Ningbo Wealth further stated to Beijing Business Today that its future hard-tech investment criteria will focus on two key dimensions: "At the industry level, we seek tracks that align with national strategic direction, offer substantial market space, and present clear, actionable commercialization pathways. At the company level, we prioritize enterprises holding core competitive positions within the supply chain and initiate positions when valuations offer reasonable cost-performance."

Xue Hongyan predicts that wealth management funds will continue to concentrate on certainty-driven opportunities in public markets, with smaller institutions participating indirectly through vehicles like funds of funds (FOFs). Wealth management firms will prioritize commercialization progress and earnings delivery capability, technological barriers and supply chain standing, and valuation-liquidity fit, while also factoring in industrial capital and state-owned shareholder backgrounds as risk mitigation references.