① The investment logic of AI has shifted towards realization, with computing power and applications advancing in parallel; ② Semiconductors are transitioning from domestic substitution to independent innovation and ecosystem upgrades; ③ Emerging industries such as innovative drugs and robotics are accelerating towards industrial validation.
Cailian Press, April 10 (reported by Wu Yuqi) Recently, Nuuo Fund released the ‘China Technology – Ignite! 2026 Tech Investment Report,’ combining insights from its research team to provide a focused analysis on AI, computing power infrastructure, semiconductors, and future industries. Compared to last year’s keyword ‘dare,’ this time the market discussion focuses not just on technological breakthroughs themselves but rather on which directions in China’s technology sector, after a phase of high-intensity iteration, are transitioning from conceptual development to industrial realization, and which segments can truly withstand the validation of performance, orders, and capital expenditures.
Behind this shift lies a subtle transformation in the narrative of tech investment. Over the past year, market attention around areas such as AI, domestic computing power, innovative drugs, humanoid robots, and space economy has continued to rise, repeatedly making the technology sector a focal point for trading capital. On one hand, the computing power chain is continuously expanding while applications are being rapidly deployed; on the other hand, valuation disagreements, bubble controversies, and crowded trading have emerged simultaneously. The enthusiasm for tech investment has not diminished but has instead entered a new phase where emphasis is placed on realization timelines and industrial sequencing.
Based on the information disclosed in this report and at the press conference, Nuuo Fund appears to emphasize a more comprehensive framework for tech investment: AI is no longer just about model competitions but is progressively unfolding along three lines—applications, infrastructure, and autonomous controllability. Future industries are no longer confined to ‘visionary narratives’ but are beginning to show more concrete industrial pathways in areas like innovative drugs, robotics, brain-computer interfaces, space economy, and controlled nuclear fusion. For public fund institutions, this shift from thematic chasing to mainline screening is redefining how tech investments are expressed.
AI is moving from technical demonstrations to commercial testing, with computing power and applications starting to accelerate in tandem.
In this wave of tech investment, AI remains the strongest driving force, but its internal logic has undergone significant changes. In the early stages, markets primarily traded based on model capabilities, parameter scales, and imaginative potential. Now, however, AI is increasingly being re-evaluated within a commercial context.
Reflecting this change, Nuuo Fund’s technology team has adopted a more pragmatic approach toward AI applications. Researcher Fan Junyi from Nuuo Fund’s tech team focuses on four key themes: AI for Business, AI for Science, intelligent assistants, and autonomous driving.
Particularly in AI for Business and AI for Science, the former corresponds to lowering professional labor barriers, while the latter reflects changes in research paradigms. Both directions point to a more practical proposition: Can AI genuinely integrate into enterprise processes, R&D workflows, and production systems rather than remaining at the demonstration level? The report mentions that some Chinese tech companies have already shortened preclinical R&D cycles by approximately 70%.
Similar transformations are occurring in autonomous driving and intelligent assistants. Autonomous driving is no longer just an old stock market concept but is entering a more intensive phase of industrial validation as highway NOA penetration exceeds 40% and urban NOA surpasses 10%. Intelligent assistants, driven by advancements in Agent architecture and RAG technology, are evolving from simple question-answering to task decomposition and closed-loop execution, signaling AI’s shift from a ‘tool’ to a ‘collaborative hub.’
Fund manager Chen Yanpeng explicitly stated that one of the confirmed directions for the next 1 to 2 years remains AI computing power. Additionally, 2026 may become the inaugural year for AI applications, with robotics, autonomous driving, and AI agents serving as critical breakthrough points.
In tandem with the rising adoption of applications, the repricing of underlying infrastructure is also gaining momentum. The report breaks down AI infrastructure into three key components: computing power, communication, and storage, offering a straightforward assessment: industry competition is shifting from ‘producing the best chips’ to ‘delivering the cheapest Tokens.’ This reflects a broader shift in market trading logic. Previously, investment focus was primarily on chip performance and model parameters, but now there is a growing emphasis on unit cost, inference efficiency, and system synergy.
Against this backdrop, general-purpose GPUs and specialized ASICs are transitioning from pure competition to a state of co-opetition, while the role of CPUs is being reassessed in the era of AI agents. Other areas such as data center interconnection, 1.6T optical modules, silicon photonics, CPO/NPO, OCS, PCB, and CXL memory pooling are increasingly being evaluated within the same industrial chain. A notable trend is that market understanding of AI infrastructure has expanded from individual devices or components to a systems engineering approach involving the coordinated evolution of ‘computing power, transport capacity, and storage capability.’
Researcher Yuan Wenchao’s emphasis on the three core components of ‘chip, network, and storage’ essentially underscores that opportunities in AI infrastructure are no longer confined to a single hardware category but rather reflect a synchronized upgrade across the entire foundational architecture.
Deng Xinyi, Deputy General Manager of the Equity Business Division and General Manager of the Research Department at Nuodan Fund, expressed optimism about the future, stating that investment opportunities in the AI sector have become clearer, more focused, and increasingly capable of delivering tangible results. The evolving landscape of model layers, alignment of demand in computing power layers, and the creation of productivity and interaction paradigms in application layers represent the most critical and enduring themes for long-term technology investment.
The semiconductor narrative is shifting from ‘substitution’ to ‘self-reliance,’ with the main technology focus extending deeper into the industrial chain.
While AI applications and infrastructure represent the hottest aspect of tech investment, semiconductors are emerging as the most foundational element in this wave of technological advancement. In contrast to the common ‘domestic substitution’ narrative, the report places greater emphasis on ‘proactive innovation’ and systematic capability enhancement.
According to the report’s findings, driven by policies, capital, and talent, China’s semiconductor industry has significantly strengthened its autonomous capabilities. Stable mass production has been achieved for mature processes at 28nm and above, advanced processes have entered scaled production, and breakthroughs continue in critical equipment such as etching, cleaning, thin-film deposition, and core materials. Advanced packaging technologies are also accelerating.
At the press conference, fund manager Zhou Jingxiang summarized this logic succinctly. He argued that self-reliance and controllability are not merely about domestic substitution but rather signify a systemic improvement in industrial capabilities. From an investment perspective, semiconductors are entering a supercycle driven by AI.
The report highlights that the strategy of ‘national chips + national models + national applications’ is accelerating, with domestic AI chips transitioning from ‘usable’ to ‘high-performing.’ For instance, DeepSeek has promoted FP8 precision adaptation for domestic chips, and Zhipu has released a state-of-the-art multimodal model fully trained on domestic computing power.
From innovative pharmaceuticals to space economies, emerging industries are no longer just distant visions but are becoming tangible realities.
In addition to AI and semiconductors, the report also dedicates significant attention to future-oriented industries.
Humanoid robots are one of the highly discussed areas. The report suggests that humanoid robots are crossing the ‘valley of death,’ with large AI models providing the ‘brain’ and the new energy vehicle supply chain offering the foundation for mass production. After Tesla Optimus accelerates its mass production, China’s supply chain is expected to secure a core share due to cost and responsiveness advantages. Yang Jingkang, an assistant fund manager, noted that the ‘dexterous hand’ represents the ‘ultimate interface’ connecting AI with the physical world and is becoming the focal point of differentiated competition.
Innovative pharmaceuticals represent another frequently mentioned focus. According to the report, over the past decade, the total value of pharmaceutical transactions in China has surged from $3.1 billion to $135.7 billion, accounting for nearly half of the global total. China has become the second-largest source of innovation for the world’s top 20 multinational pharmaceutical companies. As key products enter the approval process in Europe and the United States, China’s innovative pharmaceutical industry is transitioning from ‘going global by leveraging others’ to ‘going global independently.’ This shift reflects not only an emotional upgrade in industrial capability but also a revaluation of business models, revenue sources, and internationalization capacity.
Brain-computer interfaces, space economy, and controlled nuclear fusion are more oriented toward the medium to long term. Fund manager Tang Chen believes that brain-computer interfaces may first find applications in healthcare and could progress along the path of serious medical use, consumer-grade adoption, and eventual human-machine symbiosis. Regarding the space economy, Yang Jingkang assessed it as a ‘land grab’ crucial to national strategic security and economic leadership over the next 50 years. Liu Xiaofei added that controlled nuclear fusion has already established a complete supply chain system, with domestic enterprises continuously achieving breakthroughs and expanding international cooperation.
