The dawn of 2025 marks a pivotal transition in the realm of artificial intelligence (AI), particularly in the stock market where a noteworthy differentiation is emerging between high- and low-performing assetsThe trends observed are indicative of a more nuanced separation between various AI-related stocks and the overall market sentiment.
Last week, the AI chip giant Cambricon experienced a volatile session, with declines characterized by four out of five trading daysOn January 16, a sharp drop of 14.65% in its stock price pulled down the entire semiconductor sectorConversely, previously underperforming stocks, particularly CPO optical modules, witnessed a significant turnaroundFor instance, on January 16, Tianfu Communication surged to its daily limit with a 20% increase, while other formerly stagnant corporations like Sensing Technology and Semiconductor Manufacturing International Corp capitalized on this momentum, posting considerable gains.
It's noteworthy to mention that a recent analysis conducted by journalists at Securities Times revealed that several funds heavily invested in AI stocks are showing a marked discrepancy between their net asset values and estimated fluctuations
Advertisements
It appears that some fund managers are proactively rebalancing their portfolios ahead of market corrections influenced by overpriced AI stocks.
A Shift in AI Stock Dynamics
After reaching an extraordinary peak of 777.77 yuan—a clear testament to its market dominance—Cambricon, valued at over 300 billion yuan, is beginning to feel the frost of its lofty perchWhile the company projected promising earnings last week, its stock plummeted by a staggering 18.9%, with a resultant dip in Hygon Information by 10.17% as well.
In stark contrast to this decline, previously weak AI-concept stocks have begun to rally significantlyCPO optical modules jumped on January 16, with Tianfu Communication hitting a 20% increase that dayFurthermore, on January 17, the chip market rebounded strongly; for instance, Sensing Technology and Semiconductor Manufacturing International Corp accomplished increases of 20% and 6.31%, respectively.
This pattern is a significant departure from the past six months of market trends
Advertisements
Since the third quarter of last year, Cambricon had consistently climbed, its stock doubling amidst an environment where CPO optical modules remained relatively inert, with Tianfu Communication only slightly increasing by 8.62%. New Yi Sheng and Zhongji Xuchuang faced declines of 8.11% and a troubling 24.51%, respectively, showcasing a broader unpredictability in the chip sector as not all companies mimicked Cambricon’s consistent growth; both Sensing Technology and Semiconductor Manufacturing International Corp saw notable pullbacks of 40.03% and 25.84% at one point.
Moreover, the shifting capital flows have been starkly noticeableData from Wind revealed that last week, Tianfu Communication and Semiconductor Manufacturing International Corp attracted a net inflow of funds amounting to 2.309 billion yuan and 1.787 billion yuan, while Cambricon and Hygon Information confronted net sell-offs totaling 3.176 billion yuan and 1.43 billion yuan, respectively.
Early Adjustments by Fund Managers
The oscillation between high and low-performing stocks in the AI sphere has prompted several funds heavily invested in AI to showcase significant deviations between their net values and estimated market changes, hinting at preemptive adjustments in their portfolios.
For instance, the Wanjiay Quality Life A fund managed by Mo Haibo ranked Cambricon and Hygon Information as its top two holdings in last year's third-quarter report
Advertisements
However, on January 16, as Cambricon and Hygon Information dropped by 14.65% and 9.78% respectively, New Yi Sheng and Tianfu Communication managed to rally with gains of 6.25% and 20%. Despite Wind estimating a decline in the fund’s net value of 1.86% that day, it surprisingly rose by 1.93% instead.
Similarly, this trend was mirrored in other funds managed by Mo Haibo, such as Wanjiay Value Advantage for One Year and Wanjiay Social Responsibility, which also reported significant deviations between their net asset values and estimated fluctuationsThese discrepancies likely indicate that Mo Haibo has engaged in internal adjustments within his AI holdings, possibly reducing exposure to overbought stocks like Cambricon and Hygon Information while increasing stakes in Tianfu Communication and other performers.
On the other hand, Zhou Siyue's Dongfang Theme Selected Fund listed Hygon Information and Cambricon as the first two major holdings in the third-quarter report, yet on January 16, with both stocks experiencing significant declines, the Wind estimates forecasted a -3.9% change for the day; nevertheless, the fund increased slightly by 0.12%, pointing towards possible rebalancing as well.
Another fund to note is the Debang Xinxing Value A, managed by Lei Tao, which is the sole AI-focused fund to publish its fourth-quarter report for 2024. This report revealed a significant increase in exposure to Cambricon, marking it as the first major holding of the fund while simultaneously reducing its position in Tianfu Communication from the top ten holdings.
On the day of Cambricon's drastic drop and Tianfu Communication's remarkable rise, Debang Xinxing Value A managed a 1.32% increase, differing from Wind’s predicted 0.98% rise, suggesting that it too has engaged in minor adjustments since the beginning of 2025.
Expanding AI Trends Ahead
The variations in the AI sector, driven by both structural changes and market psychology, indicate a potential ripple effect that many fund managers anticipate will continue to unfold
- Major Western Stock Markets Strengthen
- Market Recovery: Timing Your Entry
- Global Central Banks Persist in Cutting Interest Rates
- Gold and Oil: A Bullish Outlook
- New Capital Flows into A-shares
Expectations are that the dynamics of AI trading will evolve beyond just AI chips and optical modules into hardware innovation at the end-user level.
"The previous wave of market dynamics has revealed a clear internal differentiation among artificial intelligence assetsSome tech chains benefitting from increased capital spending by overseas cloud service providers and accelerated AI computing developments are seeing their fundamentals come to fruition, similar to a 'Davis Double' effect for sectors such as optical modules and serversSimultaneously, AI applications are still lacking such fundamentals, remaining primarily within the realm of thematic investments," states Wang Bo, manager of the Southern Technology Innovation FundHe predicts that the next market cycle may maintain focus on rising value within the computing supply chain while also turning attention towards applications to fully exploit the investment opportunities presented by the ongoing innovation boom expected in 2025.
Lei Tao elaborated in his latest fourth-quarter report, emphasizing that computing power remains a highly sought-after foundational resource amid industry competition, and the sector is witnessing rapid growth
He references the strategic developments of major technology firms worldwide, highlighting that investing in computing power is both a means to stimulate current business performance and a crucial avenue for securing future competitive advantagesIn recent quarters, there has been a notable uptick in future capital expenditure forecasts from both international and domestic tech giants, exemplified by major players such as Microsoft, Amazon, and ByteDance, emphasizing the bullish outlook for computing investments.
Delving into specific sub-sectors, Lei noted that optical modules are poised to greatly benefit from the global push towards AI infrastructure, with prior performance indications having been comprehensively validatedWith continuous upgrades in computing capabilities and a surging demand for AI inference, new technologies will spur fresh product needs, alongside the emergence of new supply chain players in China, covering areas such as CPO, AEC, power sources, and liquid cooling technologies.
As for Rai Zhiyong, the lead of the champion fund, Morgan Stanley's Digital Economy Fund, he posits that 2025 may embark on a phase of intensive innovation, thus he plans to closely monitor various intelligent terminals rooted in AI, including mobile devices, glasses, and headphones.
"The key growth points for artificial intelligence lie in its applications—a parallel can be drawn with both the personal computing and mobile Internet eras
In business-to-business (B2B) contexts, attention should be given to pioneering leaders capable of quickly embracing and leveraging AI technologies to enhance product offerings and foster customer loyaltyBy honing their products to amplify customer engagement and increase average transaction value, these firms will build formidable barriers to competitionOn the consumer end (B2C), it is anticipated that 2025 will unveil numerous innovative products like AI-enhanced glasses, headphones, children’s toys, and companionship devices for the elderlyThese products, targeting massive consumer demographics, stand a chance to become runaway successes if they deliver exceptional performance, potentially igniting enthusiasm in secondary market investmentsTherefore, both B2B and B2C sectors possess substantial opportunities, expected to be key tracking directions moving forward," concludes Lei Zhiyong.