The launch of Chinese AI startup Deepseek has prompted new enthusiasm in who will be the "winners" within the global technology ecosystem. Amidst this evolving landscape, the current volatility has underscored the concentration risk associated with mega-technology stocks worldwide. Increasingly, people are contemplating the potential shift in the dominance of US tech stocks, the expansion of Chinese technology, and other prospective opportunities within the tech industry that investors should be exploring.
Terence Tsai, portfolio manager at Fidelity International comments: “Over the past two years, the technology market has been driven by a few distinct characteristics: the rise of perceived artificial intelligence winners, investor momentum, and the fear of missing out (FOMO) as investors pursued the perceived beneficiaries of US exceptionalism. This resulted in a market dominated by the so-called "Magnificent 7." These few companies were responsible for over half of the S&P 500’s 58% return over the two years spanning 2023 and 2024. Consequently, market concentration has increased, with the Magnificent 7 accounting for approximately one-third of the S&P 500 Index at the start of the year. The current risk in technology markets lies in the tendency to extrapolate and assume that this concentrated rally will continue indefinitely.
“However, while history may not always repeat itself, it often rhymes. In the mobile era, the initial victors were the Enablers, or the 'pick and shovel' plays, such as semiconductor companies. As the technology advanced and deployment costs decreased, the top performers transitioned to the Networks category, which included infrastructure and devices. Ultimately, the companies that extracted the most value were the Innovators, who created applications and software built upon the Networks and Enablers, successfully monetising user attention and engagement.
“Today's AI cycle is likely to follow a similar trajectory, with Enablers like Nvidia emerging as early winners in the cycle. However, as AI progresses towards edge devices and the developer community discovers optimal methods to monetise AI as a tool, the focus of value capture is expected to shift towards Networks and Innovators.
“Currently, semiconductor stocks constitute one-third of the MSCI World IT Index, with companies like Nvidia and Broadcom driving a significant portion of the returns. While we appreciate the long-term growth prospects of semiconductors, it is equally important to recognise the industry's inherent cyclical nature and history of sharp downturns due to oversupply from overestimating demand. Instead of trying to pick up pennies in front of a freight train by engaging in the late cycle of AI infrastructure, there are areas within the semiconductor sector that have been underperforming over the past two years, as their budgets were redirected to support the AI infrastructure development.
“Following AI’s very own Sputnik moment, investors have started to realise that technology companies outside of the US are just as competitive. DeepSeek changed the narrative and demonstrated that China’s AI ecosystem can still be competitive globally despite a lack of access to top-of-the-line compute through innovative large language model algorithms and a revitalised domestic semiconductor ecosystem that, while not yet on par with the United States, is sufficient and innovating rapidly. US companies make up 80% of the MSCI World IT Index but the US is not 80% of the world’s innovation. Looking for technology stocks outside of the US offers diversification and an opportunity to own high quality companies at a discount.
“We are currently in an environment where returns are starting to broaden out and boots-on-the-ground diligence on individual companies can make a big difference. The first few months of the year have already shown that mid-cap companies, which were largely overlooked, have started to outperform. This is not to say that large cap technology companies should be ignored, but it is equally important to continue truffle hunting for mispriced and underappreciated companies.”
China’s quantum leap: Redefining the boundaries of innovation
In recent years, China's technology sector has been somewhat eclipsed by domestic economic challenges and geopolitical tensions abroad. Specifically, since 2022, the swift advancements in AI within the United States and the imposition of chip export bans on China have raised concerns about potential technological limitations for China, with fears that major US tech companies might outspend competitors to establish a technological monopoly. However, early 2025 was a turning point when the world was taken by surprise by a significant breakthrough from DeepSeek.
Tina Tian, portfolio manager at Fidelity International comments: “DeepSeek's models have demonstrated impressive performance by integrating reinforcement learning for post-training with chain-of-thought reasoning. Notably, these achievements come at a fraction of the cost of pioneering models, while maintaining a narrow performance gap. A significant benefit to the entire AI community is that DeepSeek's models are open-source, accessible to anyone globally. DeepSeek's success underscores the adage that necessity is the mother of innovation, highlighting China's inherent structural advantages in research and development, talent, and data, among others, to drive innovation despite chip restrictions. This has prompted a reassessment of structural investment opportunities within China's tech sector.
“In recent months, the situation has become more complex due to sweeping US tariffs. Although these tariffs are temporarily paused, they still result in higher costs for the global tech sector than before. Many Chinese tech companies have diversified their market exposure by selling more to non-US markets and expanding their production footprint outside of China, actions taken since the last trade war during Trump's first term. In the short term, companies facing new tariff measures may attempt to pass these costs onto end customers. If a company possesses strong bargaining power or operates within a resilient sector where customers are less sensitive to price increases, the pass-through process will likely be smoother. In the longer term, we are likely to see continued supply chain diversification, driven by initiatives such as "China +1," where companies establish capacities outside China, particularly for goods exported to the US. Meanwhile, tariffs are part of the broader US-China strategic competition. In strategic sectors such as semiconductors and AI, we can expect accelerated technology localisation in China.
“Along the semiconductor supply chain, self-sufficiency remains low in areas such as semiconductor equipment, advanced materials, and fabless operations. As Chinese suppliers enhance their technologies and broaden their product ranges, Chinese customers are increasingly adopting local solutions, often at the expense of global suppliers, driven by concerns over supply chain disruption. In the realm of AI, China's AI ecosystem might evolve into a more closed loop, with greater localisation of both software and hardware.
“As AI technology becomes more affordable and powerful, its adoption in China is rapidly increasing. In the realm of enterprise applications, AI has demonstrated its ability to improve digital advertising conversion rates and is being integrated into enterprise software to boost productivity and customisation. On the consumer side, various AI chatbots, such as DeepSeek, Tencent's Yuanbao, and ByteDance's Doubao, are flourishing, while AI-enhanced learning devices offer highly customised and targeted learning experiences. Additionally, utilising more sophisticated large language models to distil and train smaller models enables lower inference costs and more effective AI deployment on edge devices like PCs, smartphones, and cars.
“Overall, we observe compelling investment opportunities within China's technology sector. While the sector's broad valuation remains below historical average levels, a selective approach is still necessary. We seek opportunities in AI applications that benefit from more affordable AI, such as select software companies that now integrate AI into their offerings to enhance performance while remaining cost-effective. Domestic technology substitution in semiconductor equipment and fabless design areas is also worth monitoring. From a cyclical perspective, aside from AI, the general tech sector has experienced a relatively slow recovery. Some subsegments, such as analogue and PC stocks, have not participated in the previous AI rally and present interesting prospects.”
