On November 20, 2024, at Raffles Shenzhen, Raffles Shenzhen, Ni Huangzhong, Chairman of SCY, delivered a speech on the theme of "Long-termism and Marginal Thinking of New Storage Forces in the AI Era" at the MTS 2025 Storage Industry Trends Seminar. The seminar attracted a large number of storage industry experts, entrepreneurs and investors to discuss the future development trend of the storage industry in the AI era. Mr. Ni Huangzhong's speech not only showed the latest strategic layout of SCY, but also brought a new direction of thinking to the industry, emphasizing the importance of long-termism and marginal thinking. With Ni Huangzhong's concept as the core, China Exportsemi analyzes how storage companies can achieve long-term growth and sustainable innovation in an environment of accelerated technology iteration.
Ⅰ.Current status and challenges of the storage industry in the AI era
The explosive growth of artificial intelligence has driven the rapid expansion of global data volumes. According to IDC, the global data volume is expected to reach 175 zettabytes by 2025, almost three times the number in 2020. In the face of massive data growth, storage organizations need to provide efficient, low-cost, and scalable solutions. However, AI applications place higher demands on storage, such as lower latency, higher reliability, and real-time processing power.
Figure: Ni Huangzhong, chairman of SCY, shared that in the past two years, the global storage industry has experienced a roller coaster market
Ni Huangzhong pointed out that the main challenges in the storage industry can be summarized in the following three points:
1. Technical bottlenecks: Traditional storage architectures are no longer able to meet the requirements of AI applications for high bandwidth and high IOPS (input/output operations per second).
2. Cost pressure: The cost pressure caused by the fluctuation of chip raw material prices and technological upgrades forces enterprises to seek a balance between performance and cost.
3. Intensified market competition: The R&D investment of the world's leading storage manufacturers in the field of AI is increasing year by year, and emerging companies must find a path of differentiation in order to gain a firm foothold.
These challenges show that short-term technological breakthroughs alone are no longer enough to sustain the long-term competitiveness of storage companies. In this regard, Ni Huangzhong put forward the strategic framework of "long-termism" and "marginal thinking".
Ⅱ.Long-termism: The core strategy of the storage enterprise
Ni Huangzhong believes that long-termism is the cornerstone of storage companies to cope with the changes in the AI era. He defined long-termism as the idea that enterprises do not blindly pursue short-term interests in a rapidly changing market, but achieve long-term value creation through forward-looking layout and continuous innovation.
1. Strengthen the research and development of core technologies
Ni Huangzhong emphasized that the core technology is the "moat" for storage companies to achieve long-termism. In the AI era, the evolution of storage technologies such as NAND and DRAM and the development of next-generation storage media (such as MRAM and PRAM) will become the focus of competition. He shared his practical experience of current creativity:
- R&D investment accounts for more than 20%: 20% of the company's annual operating income is invested in core technology R&D, with a special focus on memory controllers and AI acceleration chips.
- Interdisciplinary innovation: Optimize storage system performance by introducing AI algorithms, such as intelligent compression and predictive caching techniques, to improve overall efficiency.
2. Build ecological partnerships
Long-termism also requires enterprises to go beyond the "go it alone" development model and establish in-depth cooperation with upstream and downstream enterprises, academic institutions and technology alliances.
- Collaborative innovation with cloud computing companies: We have cooperated with leading cloud computing companies at home and abroad to jointly develop storage solutions for AI training and inference.
- Participate in the formulation of industry standards: Participate in standardization organizations such as JEDEC and OCP to ensure that new technologies can be quickly industrialized.
3. Focus on sustainability
In today's era of resource scarcity and increasing environmental pressure, Ni Huangzhong pointed out that the development of green storage technology is an important part of long-termism. For example, through the development of low-power memory chips and recyclable materials, the company has significantly reduced the carbon footprint of its product life cycle.
Ⅲ.Thinking Marginally: Optimizing the Economics of Storage Systems
In addition to long-termism, Ni Huangzhong also put forward a unique concept - marginal thinking, that is, through refined management and technological optimization, to maximize resource utilization and marginal returns. This philosophy is especially important in the actual operations of storage enterprises.
1. Improve marginal performance
The storage requirements in AI applications have non-linear growth characteristics, such as from the initial data collection to the later stage of model training, the storage performance requirements will change significantly. Centicreative optimizes marginal performance by:
- Tiered storage architecture: Combines high-performance storage (such as NVMe SSDs) with high-capacity storage (such as HDDs) to achieve a dynamic balance between performance and cost.
- Intelligent O&M: AI technology is used to predict storage device failures and load peaks, reduce the risk of system downtime, and improve utilization.
2. Reduce marginal costs
In AI-driven data storage, hardware costs and energy consumption account for a major portion of enterprise spending. Ni Huangzhong shared the practice of SCY:
- Modular design: Reduce production and maintenance costs and improve system scalability through modular storage hardware.
- Energy efficiency optimization: In response to the high energy consumption problem of data centers, liquid cooling technology and dynamic power management strategies are adopted, saving energy costs by more than 15% per year.
3. Expand marginal markets
The ultimate goal of marginal thinking is to enhance the overall market competitiveness of enterprises. Ni Huangzhong emphasized that by targeting differentiated market needs, storage companies can find blue ocean opportunities in the red ocean market. For example, the company focuses on serving the customized storage needs of small and medium-sized AI enterprises and industry application scenarios (such as autonomous driving and intelligent manufacturing).
Ⅳ.The new SCY product was released
At the end of the speech, Ni Huangzhong announced the full mass production of the 1TB UFS3.1 memory chip developed and manufactured by SCY. According to Ni Huangzhong, this product is the first mass-produced 1TB storage product in Chinese mainland, marking an important progress in the field of high-end storage technology in China. The chip supports the UFS 3.1 standard, has higher read and write speeds and lower power consumption, and is especially suitable for high-end consumer electronics such as smartphones and tablets. Compared with the previous generation of UFS 3.0 products, firstly, the capacity has been increased from 512GB to 1TB, and secondly, UFS 3.1 has increased the read and write speed by more than 50%, especially in the random read and write performance. In addition, it has better energy efficiency, which can balance high performance with low power consumption, extend the battery life of the device, and exhibit better stability and durability under high load conditions. Overall, the 1TB UFS 3.1 memory chip not only improves the storage performance, but also promotes the innovation of domestic storage technology, which is of great industry significance.
Picture: Ni Huangzhong announced the full mass production of the new 1TB UFS3.1 memory chip developed and manufactured by SCY
Ⅴ.Conclusion: The future of the storage enterprise
The storage industry in the AI era is full of opportunities and challenges. In this rapidly changing market, short-term technology leadership is not enough to sustain the long-term development of enterprises. Ni Huangzhong is guided by long-termism, emphasizing the deep cultivation of core technologies, ecological cooperation and sustainable development, and at the same time opening up a differentiated development path through marginal thinking and refined management of resources.
The experience of SCY shows that only by taking a long-term view and practicing marginal thinking can storage companies ride the wind and waves in the AI wave and write their own future chapter.
